Mugnsoft logo Mugnsoft
Changelog

Core release notes

Detailed information about every release of the Mugnsoft Core monitoring platform — webserver, probes, sentinel discovery agent and integrator. Looking for the portable journey builder? See the Pulse release notes. For the installation guide, visit doc.mugnsoft.com/installation.

v4.3.0

latest

v4.3.0 adds SIGNL4 alerting, moves PagerDuty to the Events API v2 and makes the interface easier to work with day to day: a smoother 3D view, clearer messages when a component does not answer, and CSV exports that make a complete backup. It is a drop-in upgrade from 4.2.x: upgrade in the usual order (integrators, monitor probes, sentinel agents, webserver, then webserver-front). Your configuration, thresholds and alert destinations carry over unchanged.

＋ New features

  • SIGNL4 alerting. Alerts can go to a SIGNL4 team through its inbound webhook, with the same On fail and On change switches as the other channels. Each monitor raises one SIGNL4 alert, and the alert is closed automatically when the monitor is back to OK. Set a team secret on the monitor, or a default one on the integrator. Application roll-ups and sentinel agents can alert SIGNL4 too, and the setting travels with monitor exports and imports.
  • PagerDuty Events API v2. PagerDuty alerts now use the Events API v2, which replaces the deprecated v1. Each monitor keeps a single incident: later alerts are grouped into it and OK resolves it. The incident severity follows the status (CRITICAL, MAJOR / ERROR, MINOR), the alert details arrive as custom details, and a link opens the monitor in the 3D view. Your PagerDuty integration keys stay where they are.
  • Send a test notification. From the integrator settings, send a test alert on email, Slack, Teams, PagerDuty, SIGNL4 or script. It leaves from the integrator's host, the same path real alerts take. Leave the destination blank to test the saved default.
  • Status priority for URL and API monitors. When a monitor has pattern groups, choose how the pattern verdict and the response time combine: the most severe of both (the default, unchanged), patterns first, response time first, or patterns only.

▲ Improvements

  • 3D view. Nested applications are laid out as a grid of sub-applications. The weather view keeps its own layout, separate from the standard one. Only Save layout stores positions, so leaving the page no longer saves them by accident. Scrolling inside a popup no longer moves the camera, and hovering over applications is more precise.
  • Clearer component messages. Every page now says the same thing when a probe or integrator does not answer, and whether the host is unreachable, the service is down or it sent no data. Adding a component explains why the login failed (unreachable, timed out or refused), and the form keeps what you typed.
  • Complete CSV backups. Users, reports and downtimes can now be exported and imported like monitors and applications. Application logos are embedded in the file, and a user line carries that user's 3D layouts, cameras, view preferences, background and profile picture, so a profile moves between installations in one file. Imported reports and downtimes are scheduled straight away.
  • Safer imports. Re-importing an existing monitor keeps the thresholds tuned since the export. A bad row is reported with its line number and what to fix, and the rest of the file is still imported.
  • Application reports cover sub-applications at any depth. Related monitors load on their own, and every report type previews the monitors it covers.
  • Component metric charts show outages and silent periods as no data bands instead of drawing a line across them, and the monitor report charts follow the same rule.
  • The compare page ranks monitors on the right figures, and a probe-wide baseline compute reports what it updated or skipped.
  • New translations for French and Spanish across the new screens.

✕ Bug fixes

  • The web session could expire while you were browsing between pages, or after the computer woke from sleep. It is now renewed in time.
  • An alert could show the wrong duration for a fresh status, for example a new CRITICAL reading for 3 d after three days OK.
  • Behind a proxy, URL monitors now read the certificate on the check's own connection, which respects the monitor's proxy and timeout. Plain http monitors no longer raise a false certificate about to expire alert.
  • A check faster than 1 ms or slower than one minute could lose its result. Results sent to a slow integrator are no longer cut short and sent twice.
  • Probes on a filtered network can download the EUM web drivers through their outbound proxy.

⬢ Component updates

  • Probe log messages from the HTTP client now go to monitor.log with the monitor, target and proxy, and never include credentials or query strings.

v4.2.0

v4.2.0 adds WebSocket and Kubernetes / Docker monitoring, splits the SLA into availability and performance, and puts a status timeline on every application. Reports can be sent as PDF, and the interface and the reports are available in French and Spanish. Read Before you upgrade first: license usage now counts WebSocket and workload monitors, licenses are bound to a hardware fingerprint, the internal component secret is no longer shared, and the release signing key has changed.

⬢ Before you upgrade

  • License usage counts WebSocket and workload monitors. Each WebSocket monitor and each workload monitor costs 1 point, like a URL or TCP monitor. A workload monitor exists once per probe; the container host itself stays free, like SNMP and WMI devices. Over the limit nothing stops running — only the creation of monitors is refused, from the interface, the API and discovery. Count them before upgrading and ask for a license extension if needed.
  • Licenses are bound to a hardware fingerprint. A webserver or integrator license is now generated from the fingerprint of the server it runs on, which replaces the binding to an IP address. Capture it with the fingerprint tool from the download page and send us the token. A resized virtual machine keeps its license; licenses already issued keep working.
  • Set your own internal component secret. _sharedInternal is the password of adminMNS, the account the components use to drive each other, and earlier packages shipped the same value to every customer. The 4.2.0 packages ship it empty: a new webserver generates a random one on first start, and every component logs a SECURITY warning while it still uses a packaged value. There is no built-in fallback any more — put the same value in webserver.json, monitor.json, integrator.json and discovery_agent.json, then restart, or Resync tokens, Enable and Push Settings fail with an HTTP 401.
  • Upgrade from the edge inward: integrators, monitor probes, sentinel agents, webserver, then webserver-front. A 4.2.0 webserver in front of older probes shows the new monitor types but cannot create them on those probes.
  • Sentinel agent process I/O metrics changed unit. iostatRead and iostatWrite now measure per-process throughput in KB/s. Review the thresholds set on these two metrics.
  • Alert emails have a new subject, leading with the status and the key fact — for example [MAJOR] web-shop · Execution time: 580 ms · probe-paris. Mail rules that match the old Mugnsoft - … status is … subject must be updated.
  • The release signing key has changed. Archives are signed with a new key from this release on, and the archives of earlier versions have been signed again with it. Import Mugnsoft_public.key again before verifying a download — fingerprint 65F8 B880 F8CC 9DCF D2C8 4887 8FEE 1578 1BEF 9B7A.

＋ New features

  • WebSocket monitor. Opens a session, sends your messages, validates the replies and grades connect, latency and session timing, with certificate expiry and identity checks like the URL monitor.
  • Container monitoring (Kubernetes and Docker). A container host device discovers workloads from the Kubernetes apiserver or the Docker Engine API; adopt the ones to watch and each becomes a workload monitor (replicas, restarts, OOM kills, CrashLoopBackOff). Drift is reconciled: new workloads are proposed, vanished ones are kept and disabled.
  • Probe inside Kubernetes. A Helm chart and a slim probe image run a probe in the cluster it monitors.
  • SLA availability and performance. An SLA is now a target plus an impact condition, tracked separately for availability and, optionally, performance.
  • Application timelines. Status timelines in the application table, in the application report and under each application's status chip in the 3D view.
  • The interface in French and Spanish. Each user picks a language in their profile; toasts, dialogs, tables and tooltips follow it.
  • Reports in French and Spanish, as HTML or PDF. Each report has a language, followed by its page and its email, and can be sent as HTML, PDF or both. The PDF is printed by a headless Chrome, Chromium or Edge on the webserver host; when none is found the HTML report is sent instead and the send result says why.
  • Reports that open without Internet access. The webserver setting HTML report libraries can embed the chart libraries in each emailed HTML report, at about 820 KB more per file. The default is unchanged.
  • Database column checks. A database monitor can grade values of its result set, like a pattern group does for URL and API monitors. Pattern groups and column checks also accept string in / string not in, number between / number outside, and a "must not contain" content check.
  • UDP monitor modes — reachability only or request / response, so a UDP status and response time only claim what the protocol can prove.
  • Default notification destinations on the integrator. A monitor that switches a channel on and leaves its destination blank is delivered to the integrator's default; a destination set on the monitor always wins.
  • Events: service-level root cause, deletion and paused monitors. The incident strip names the service, not only the host; events can be selected and deleted together; a disabled monitor or application shows as PAUSED instead of not reporting.
  • Address changes follow through. When an integrator's address is changed in Components, probes and sentinel agents receive it, with a retry every 5 minutes for the unreachable ones, and results buffered for the old address are sent to the new one.
  • SMTP "Verify certificate" — a switch in the SMTP settings, honoured by every component. Off by default: mail servers with self-signed certificates keep working.

▲ Improvements

  • Alert notifications show the same facts on every channel (email, Slack, Teams, PagerDuty): status and for how long, target, probe and location, check time and message. Emails are colour-coded by severity and carry a plain-text part; Teams alerts are Adaptive Cards and Power Automate Workflows webhooks are supported.
  • Monitor reports gain a summary strip (share of OK checks, non-OK checks by severity, response time mean / p95 / min / max), response bars coloured by status, jitter and packet-loss panels on ping reports, and one panel per numeric pattern group on URL, API and database reports.
  • WebSocket, WMI and workload monitors have their own section in the All and application reports, and a downtime can cover WebSocket and workload monitors.
  • Scheduled reports look like the report shown in the interface, EUM, application and All reports included.
  • Backups copy every database of a component into one folder per run, readable by the service account only; Backup all requires the admin role.
  • Linux packages extract with their programs executable (binaries, scripts, the bundled JDK and ffmpeg): no more chmod +x after unpacking.
  • New tags are stored in lower case; tags that already exist keep their spelling.

✕ Bug fixes

  • A URL monitor whose request failed outright (connection refused, timeout, DNS or certificate error) showed ERROR on its own page but never reached the integrator: no event, no notification. A URL monitor with an unusable proxy configuration was not reported at all. Both are now reported like any other result — upgrade your probes to get this fix.
  • A failed URL or API monitor now gives the cause — the request error, or the HTTP status that came back, for example The API answered HTTP 500 Internal Server Error. — instead of Failed to reach the API. Mail rules that match the old text need updating.
  • REST API: POST /api/apis failed with a bare 400 whenever headers was left out or held real headers. Headers are now optional and accepted as a JSON object or as the documented array of objects.
  • API monitor export wrote request headers in a form that could not always be imported back; it now writes a JSON object, and the import still accepts the older list.
  • URL and API monitors created through the REST API with insecureCert stored a value the probe ignored. Such a monitor created with insecureCert: true starts skipping certificate verification after the upgrade, as it was asked to.
  • Exporting system checks failed; SNMP v3 devices using SHA-224 to SHA-512 or AES-192/256 could be exported but not imported back; a host import refused notifyStatus written in lower case.
  • The events page could mark the whole estate not reporting when one integrator output was slow.
  • Timelines showed spurious unknown bars.
  • UDP reports were never sent, and ping reports showed microsecond values labelled as milliseconds.
  • ADMIN to webserver-front replication: an older copy could overwrite a newer one, deletions did not always reach the front, and a deleted user's session stayed valid.
  • EUM reports drew a new series each time a script was edited: steps are now merged by name.
  • Backups skipped the SNMP / WMI monitor configurations and container hosts on the webserver, and the log and folder monitoring history on the sentinel agent.
  • The sentinel agent's /docs page showed the webserver API instead of its own.
  • Pre-release audit fixes across all components, notably alerting, log and folder monitoring.

⬢ Component updates

  • All components are built with Go 1.27.2 and updated dependencies. At release, govulncheck reports no known vulnerability reachable from any component.
  • The adminMNS machine account no longer ships with a shared password, and the webserver-front never carries the account.
  • License fingerprint tool v1.0.0 for Linux and Windows, signed like the archives, on the download page.

v4.1.1

v4.1.1 improves interface clarity and consistency, making navigation and data analysis easier in everyday use.

▲ Improvements

  • Improved report presentation for clearer and easier reading.
  • Clearer and more consistent navigation between views.
  • Enhanced tooltips providing more context and useful information.
  • A more consistent interface with harmonized spacing, table layouts, and active navigation styles across the application.

v4.1.0

v4.1.0 makes discovery protocol-aware: every captured connection is now labelled with the application protocol or framework it actually carries — gRPC, WebSocket, Arrow Flight SQL, HTTP/2, PostgreSQL, Redis, Kafka, MQTT — together with the signal the verdict came from and how confident it is. It also puts an end to configuring blind: SNMP OIDs and WMI queries are proposed from a catalogue and verified live on the target through the probe, log and folder rules can be tried against the real file, and downtime windows, SLA exclusions and report scopes all show what they will affect before they are committed.

＋ New features

  • Application-protocol identification on the discovery map. Flow edges now carry the protocol or framework they transport. Three independent signals feed the verdict: signatures in the opening bytes of a flow (HTTP/1.x, HTTP/2, gRPC, Arrow Flight SQL, WebSocket, PostgreSQL, MySQL, Redis, AMQP, Kafka, MongoDB, MQTT, SSH), TLS handshake metadata (ALPN and SNI, always cleartext), and process/port inference as a last resort. Every label states its source (dpi, alpn, process, port) and its confidence (observed, inferred, guessed), so an inferred protocol is never read as a measured one. The map can be filtered by protocol, and detection can be switched off per agent.
  • SNMP OID and WMI query assistants. Both device types propose metrics from a catalogue keyed by device type, verify the whole selection in bulk against the real target through the probe, and let you save reusable named templates. An OID walk browser and a WMI class browser are available for anything the catalogue does not cover.
  • Downtime calendar and impact preview. Downtime windows can be browsed as a month calendar next to the existing table, with a range picker for the period. Before saving, an impact preview lists exactly which monitors and applications the window will silence.
  • SLA exclusion windows. Maintenance and known-issue periods can be excluded from SLA uptime with a documented reason. Outages inside a window no longer count against the SLA, and corrected periods are flagged with a ★ in the SLA history.
  • Report scope preview. Every report type can resolve its tag, application and probe selection into the actual list of monitors it will cover, on demand, from the report form.
  • Threshold provenance. For each system and process metric the interface states whether the threshold that governed the run came from the auto baseline or from the fixed fallback — a fixed value while auto is enabled means the baseline was unavailable. Fixed thresholds are editable per metric, and baselines can be recomputed on demand.
  • Capture availability is reported, not guessed. When a host has no packet-capture runtime, the topology says so explicitly instead of showing zero traffic that looks like an idle host. A merged view only claims capture when every contributing agent had it.
  • Log and folder monitor test. Log rules and directory rules can be run against the real file or folder from the interface, and their matches inspected, before the monitor is saved.
  • Several named pattern groups per URL and API monitor, evaluated to a worst-status verdict, with contains checks and request bodies typed from their content type.
  • URL and API monitor test panel — a monitor can be run on demand from its form and the resulting steps inspected, alongside a rewritten inline HAR viewer.
  • CPU iowait and TCP socket count added to system monitoring on the probe side.

▲ Improvements

  • Alert notifications are unified: e-mail, Slack, Teams and PagerDuty messages all carry the monitor type, the originating probe and a status colour, and look identical whether they were sent directly by a probe or routed through the integrator.
  • The consolidated monitor listing gains a multi-select type filter, paging controls moved into the table header, and a layout consistent across the monitor, application and server tables.
  • Loopback traffic is collapsed into a single edge per flow, with direction taken from the listening socket; loopback capture itself is toggleable per agent.
  • Discovery thresholds treat an empty value as unset everywhere, and report validation errors and warnings when a configuration is saved.
  • Stricter date and time validation on URL and API pattern checks.
  • License validation is cached instead of re-read on hot paths, and logs a single line rather than repeating itself.
  • Scheduled jobs are panic-protected on both webserver and probe: a failing job no longer takes the scheduler down with it.
  • Repository backup supports a self-managed GitLab base URL, with inline setup help for both GitLab and GitHub.
  • The component configuration builder introduced in v3.2.0 now covers URL and API monitors, log rules, folder rules, traffic capture and notification routing, and the generated JSON can be copied as well as downloaded.
  • Interface polish: synchronised chart hovering and save-as, a shared spinner component, stacked off-canvas panels, and discovery views served on the read-only DMZ front server.
  • Training material (Viewer, User and Administrator decks, English and French) refreshed for this release.

✕ Bug fixes

  • Sentinel agent threshold computation: flat series are skipped, empty performance windows no longer produce absurd baselines, and a minimum sample count is required before a baseline is trusted.
  • Loopback process monitoring attributed traffic to the wrong process.
  • Discovery agent thresholds were not reflected in the details panel.
  • Integrator external-services connectivity check.
  • Directory metrics: file counting, recursive counting and total size corrected, including for missing directories and disabled rules.
  • Pattern-group metric exposure, export and import of URL and API monitors, and table layout and filter regressions on the consolidated listing.

⬢ Component updates

  • Sentinel discovery agent — protocol identification with per-flow caching; log monitoring hardened against rotation by truncation and by rename, with multi-line joins, captured line samples, throughput burst/silence and staleness checks; directory metrics reworked; loopback edge collapsing; capture status reporting.
  • Monitoring probe — bulk OID test and walk, WMI query test and class browsing behind the configuration assistants; content-type-aware API request bodies; date/time validation on pattern checks; auto-threshold computation reworked; probe-owned process detection to clean up orphaned browser and CLI processes; Linux install/uninstall scripts and concurrency rework.
  • Integrator — external-services connectivity check corrected, and alert payloads aligned with the new notification format.
  • Predictive analytics — shipped as a standalone packaged runtime, with consistent window and horizon resolution, and early stopping enabled only when the time range actually yields a validation batch (with an explicit warning when it does not).

v4.0.0

v4.0.0 is a major release on three fronts. SNMP and WMI devices become first-class monitored objects, with auto-discovery and their own place on the topology. A read-only replica of the webserver can now be published in a DMZ, fed by one-way mTLS replication, so status can be exposed outside the trusted network without exposing the platform. And identity moves to single sign-on, with ITSM ticketing wired directly to monitor state.

＋ New features

  • WMI device monitoring. Windows hosts can be monitored through WMI queries, with per-device status, a dedicated device table, an actions tab and a split detail pane.
  • SNMP devices. Remote devices become SNMP devices as a monitored object in their own right, with SNMP auto-discovery, a global status column and their own representation on the topology map.
  • Read-only DMZ front server. A second webserver can be deployed in a DMZ and fed by one-way mTLS replication from the trusted webserver. It serves a read-only view, requires no license of its own, and applies a configurable status grace period so brief transitions never reach the public view.
  • OIDC single sign-on — a third authentication method alongside local accounts and LDAP.
  • ITSM ticketing. ServiceNow, GLPI and Jira tickets are created and closed from monitor state, with resolution deferring to the ITSM side so tickets already owned by a human are not closed automatically.
  • Heavy-client (GUI) user journeys gain replay support, and EUM alerts now carry per-transaction breach information.
  • One consolidated monitor listing. The eleven per-type monitor pages are replaced by a single listing covering every monitor type, with per-type filtering.
  • 3D application view: friendly view and sharing. A simplified "weather" rendering for non-technical audiences, per-user saved layout and camera, and the whole view embeddable in an iframe.
  • Filesystem inode usage monitoring, application-level downtime, a status legend, reports as off-canvas panels, and user profile pictures.
  • Training material: Viewer, User and Administrator decks, in English and French.

▲ Improvements

  • Device pages reworked: split detail pane, actions tab, icons in the device table, and monitor updates that keep collected data.
  • Status data is purged when the monitors that produced it are removed.
  • Performance and safety: 3D scene teardown, XSS hardening, bounded concurrency in file cleanup paths, and a data race fixed in bandwidth accumulation.
  • Consistent page spacing, table layout and active-navigation styling across the interface.

✕ Bug fixes

  • SNMP monitors were not scheduled at startup; SNMP device updates and UDP monitors corrected.
  • API monitor import.
  • Memory leaks on the SNMP device page and in the 3D application view.
  • Application graph expansion, sub-applications on the front server, 3D banner scrolling, and a device refresh loop.

⬢ Component updates

  • Webserver front — new component: read-only DMZ replica, no license required, one-way mTLS replication, configurable status grace period.
  • Monitoring probe — WMI support, SNMP startup scheduling fix, purge of stale status data, memory-leak fixes, bounded concurrency in cleanup paths, GUI journey replay and per-transaction EUM breach reporting.
  • Sentinel discovery agent — filesystem inode monitoring, additional system metrics, and a data race in bandwidth accumulation fixed.

v3.3.0

v3.3.0 adds the two things that turn monitoring into reporting: SLA figures you can defend, and forecasting that shows a metric heading for its threshold before it gets there. The discovery agent also learns to read log files.

＋ New features

  • SLA reporting. Availability is computed and reported as an SLA, with weighting per monitored element and manual corrections for periods that should not count as measured.
  • Forecasting. Every monitor type can carry prediction fields, served by the predictive analytics engine, so a metric trending towards its threshold can be seen before it crosses it.
  • Log file monitoring on the discovery agent. Pattern rules over log files with alerting, read offsets persisted across restarts, and a bounded number of lines read per tick. Database connection counts are monitored and alerted on the same way.
  • Automatic baselines. Global system metrics get auto-computed baselines on the agent, and ping jitter gets auto-computed thresholds on the probe, reported through the API alongside the measured values.

▲ Improvements

  • Missing-process alerting follows the same notifyAfter and notifyFor semantics as every other alert.
  • Discovery agent configuration: compact threshold tables and traffic toggles instead of long forms.
  • Report presentation: timeline labels, pie chart sizing, and the visibility and capture controls moved side by side.
  • Regex caching and retention purging on the agent, with a configurable cap on lines read per tick.

✕ Bug fixes

  • Jitter standard deviation overflowed on large values, and is now computed as a float.
  • ICMP TTL handling reverted on Windows raw sockets, where it is unsupported.
  • Agent registration panic and administrative authentication failures.
  • Status colour of MINOR events in the event browser.

⬢ Component updates

  • Sentinel discovery agent — log file monitoring with alerting and offset persistence, database connection monitoring, auto-baseline thresholds for global system metrics, regex caching and retention purge.
  • Monitoring probe — jitter thresholds with auto-computation and API reporting, prediction fields on every monitor query type.
  • Predictive analytics — InfluxDB v3 support, a larger LSTM, early stopping, configurable window size and tutorial scripts.

v3.2.0

v3.2.0 is about finding things in a large estate. Every map and listing gains a search bar, the 3D application view remembers how each user left it, and network traffic becomes a first-class threshold rather than a number to read off a chart.

＋ New features

  • Search everywhere — a search bar in the infrastructure topology map, the 3D application view, the application view and the monitor listing, so an element can be located without knowing where it sits.
  • Traffic thresholds. Inbound and outbound traffic can be given thresholds with a configurable averaging window, so a short burst does not raise an alert but a sustained one does.
  • A table view of the consolidated monitor listing, with multi-select deletion, next to the existing card view.
  • Component configuration builder. A guided, question-by-question assistant produces the configuration files for the probe, the integrator and the Sentinel agent, with a reset available at any point and the resulting JSON offered for download.
  • New discovery agent system metrics: iowait, CPU user and others, charted alongside the existing ones.

▲ Improvements

  • 3D application view: layout and camera saved per user, per-user background image, status banner, help modal and a summary bar.
  • Infrastructure and server topology graphs redrawn, with external services shown in the component topology and interface names carried on the edges.
  • Failed notifications are buffered to disk and retried, on webserver, probe and agent alike.
  • Alert routing gains notifyAfter and notifyFor semantics.
  • Schedule editor overhaul: day toggle pills, a live summary of the resulting schedule, and validity checks.
  • Security-scan and memory-leak sweeps across webserver, probe and agent, with allocation reductions in the Go code.

✕ Bug fixes

  • Zoom synchronisation across charts did not complete on every chart.
  • Edge state was wrong when a node was disabled.
  • Server status refresh interval in the 3D view.
  • Discovery agent URL and API lists.

⬢ Component updates

  • Sentinel discovery agent — traffic thresholds with averaging window, new system metrics, richer edge information, memory and performance fixes.
  • Monitoring probe — disk-backed notification retry buffer, allocation reductions and security-scan fixes.
  • Integrator — disk-backed retry buffer for failed notification posts.

v3.1.0

v3.1.0 rebuilds SNMP monitoring on the probe and teaches the discovery map to report volume, not just shape: network traffic in and out is now accounted per process and rendered on the map.

＋ New features

  • SNMP monitoring rebuilt. SNMP becomes its own module on the probe, with reworked scheduling and result handling, and a configuration and results view to match in the interface.
  • Per-process network traffic. The discovery agent accounts inbound and outbound traffic for each discovered process, and the discovery map renders it alongside the existing flows.

▲ Improvements

  • Discovery map nodes and edges carry richer information, and the agent configuration layout was reworked around it.
  • Component configuration is transported encoded between webserver and probe, so no character in a monitor definition can break a configuration push.
  • Front-end memory leaks fixed and shipped web assets trimmed.

⬢ Component updates

  • Monitoring probe — SNMP extracted into a dedicated module, monitor configuration transport reworked, web-driver handling simplified.
  • Sentinel discovery agent — inbound and outbound traffic accounting per process.

v3.0.0

The v3.0.0 release is about seeing the whole estate at once. Applications and their monitors are laid out in a navigable 3D view, and the discovery agent moves from process inventory to a full status story: a timeline split by scope, a graph that can be replayed over time, per-process system and performance metrics, and the network flows between processes.

＋ New features

  • 3D view of applications and their monitors: each application is a plane holding its monitors, colour-coded by status, with nested sub-applications, hover details (type, probe, status, tags) and a camera position and layout saved per user. 3D application view 3D view of the estate
  • Discovery status timeline, split into global, system and process scopes, with the number of state changes and the time spent degraded over the selected period. Discovery status timeline and graph
  • Network flows between discovered processes are rendered on the discovery graph with their traffic volume, next to per-process CPU, memory, swap and I/O and the host's own system metrics. Network flows between processes Network flows between processes

▲ Improvements

  • The discovery graph is explorable by depth level, collapsible, and can be replayed at an earlier point in time rather than only showing the current state.
  • A single agent configuration file now carries everything the agent needs: monitored processes and ports, filesystems, per-metric automatic and fixed thresholds, the URL and API monitors it owns, and its notification routing. Discovery agent configuration

✕ Bug fixes

  • Discovery charts failed to render for metrics with no data over the selected period.

v2.6.0

The v2.6.0 update brings a comprehensive overhaul to the Mugnsoft core, featuring significant improvements in Web Application Performance Monitoring with the discovery agent. This agent simplifies access to system and performance metrics of a host server and monitored processes by identifying these processes and their metrics, including interprocess communication. All monitored elements are now consolidated on a single web page, making performance monitoring more efficient. These updates aim to enhance the monitoring of web business application performance. With business propagation rules, users can gain better insights into how each monitored element affects the application's performance and availability.

▲ Improvements

  • Import from CSV files gets simpler. CSV file import
  • Enhancement of the Web User Interface to achieve a straightforward yet effective user experience.
  • Alerts are now centralized within the Integrator component, simplifying network flow operations. Centralized alerting Centralized alerting

✕ Bug fixes

  • Minor bugs fixes.

＋ New features

  • The previous release laid the groundwork for the discovery agent, which is responsible for identifying processes, their system and performance metrics, as well as interprocess communication. This release officially introduces the discovery agent. Discovery agent monitoring Discovery agent monitoring Remote host monitoring Network flows between processes are automatically discovered, and system and performance metric thresholds can be configured automatically.
  • An Event Browser is available to show current alerts. Event browser
  • Application statuses are now better organized, providing a clear overview of everything at a glance.
  • Application statuses can now be publicly shared and embedded on a web page. Embedded application statuses Embedded application statuses

v2.5.0

The v2.5.0 release introduces remote script and binary execution, enabling you to monitor anything without installing an agent. This update also includes minor bug fixes. Additionally, it lays the groundwork for a discovery agent, which will be introduced in the next release.

▲ Improvements

  • Better handling of import of CSV files to create batch of monitors. CSV file import
  • This release also lays the groundwork for the discovery agent, which will be responsible for identifying processes, their system and performance metrics, as well as interprocess communication.

✕ Bug fixes

  • Headers for the API monitor are now correctly laid out. API headers

＋ New features

  • You can monitor anything you want without the need to install an agent by defining remote hosts and uploading any necessary scripts or binaries. Remote hosts can be managed by one or more monitoring probes. Remote host monitoring Remote host monitoring Remote host monitoring Remote host monitoring When defining a remote host, 2 monitors (tcp and ping) of type SYS will be created automatically. The state (enabled/disabled) of these 2 monitors are the state of the remote host itself.

v2.4.0

The v2.4.0 release introduces a groundbreaking feature: heavy client user journey monitoring. What does this mean exactly? You can now use Mugnsoft to monitor large, rich client applications (not web-based but IDE-like). With the release of Zabbix 7.0 in June, Mugnsoft already provides support for this Zabbix version.

▲ Improvements

  • It is possible to connect to Zabbix API through the use of a Zabbix API's token. Zabbix token connect
  • It is possible to connect to the latest version 7.0 of Zabbix. Zabbix v7
  • It is possible to schedule regular downtime windows. Schedule regular downtimes

✕ Bug fixes

  • Some historical charts displays fails when there is no data.

＋ New features

  • Introducing heavy client user journey monitoring! Mugnsoft now allows you to monitor user journeys on large, rich client applications (IDE-like, non-web-based). Track and analyze the interactions of your users in real-time to improve performance and user satisfaction. Heavy client monitoring

v2.3.0

The v2.3.0 brings minor improvements and new features detailed below.

▲ Improvements

  • It is possible to connect to Zabbix API through the use of a Zabbix API's token. Zabbix token connect
  • It is possible to schedule regular downtime windows. Schedule regular downtimes

✕ Bug fixes

  • N/A

＋ New features

  • It is now possible to get username and password from secrets in a Key Vault.
    Currently supported Key Vaults are:
    • Hashicorp Vault Hashicorp Vault
    • Azure Key Vault Azure Key Vault
    • Cyberark Conjur Vault Cyberark Conjur Vault
    Mugnsoft IDE Key Vault Mugnsoft Webserver Key Vault Mugnsoft Monitor Key Vault

v2.2.0

The v2.2.0 brings minor bugs corrections.

✕ Bug fixes

  • Application color adjustment in service tree.
  • Application: state and status of the top app are not correctly reflected in service tree.
  • Multi-tenant: users with an app defined in their profile could not always see the app in the app page.

v2.1.0

The v2.1.0 brings 1 new feature: the possibility to record scenarios including TOTP MFA. And 1 improvement: Mugnsoft alerts in Zabbix are more descriptive.

▲ Improvements

  • Alerts in Zabbix are now more descriptive. More details on Mugnsoft event in Zabbix

＋ New features

  • It is now possible to record and replay scenarios with TOTP MFA enabled. TOTP MFA

v2.0.0

The v2.0.0 introduces enhancements to the Web Application Performance Monitoring user interface. This update enables filtering comprehensive monitoring of critical web application elements, such as TCP connection time, DNS lookup, remote metrics check, and database query performance. All monitored elements are now conveniently accessible on a single web page, streamlining the performance monitoring process. These changes are designed to enhance the monitoring of web business application performance. By using business propagation rules, users can better understand the impact of each monitored element on the application's performance and availability.

▲ Improvements

  • Cosmetic improvements on the user interface. Header navigation menu has been reorganized. Layout
  • End User Monitoring metrics report has been redesigned. Layout

＋ New features

  • Filtering of items of critical web application elements, such as TCP connection time, DNS lookup, remote metrics check and database query performance. New monitored item
  • All monitored elements are now conveniently accessible on a single web page, streamlining the performance monitoring process. New monitored item
  • Application definition and performance monitoring. By using business propagation rules, users can better understand the impact of each monitored element on the application's performance and availability. Applications Applications

v1.8.0

On the Mugnsoft core side, this v1.8.0 release brings notable enhancements to the web user interface. All components are now organized within a single web page, and web scenarios can be automatically deployed based on probe utilization. The auto-deployment algorithm selects the least used probes, while components include system metrics usage charts for CPU, memory, and load average comparisons.

▲ Improvements

  • Cosmetic improvements on the user interface. Header navigation menu has been reorganized. And components have been moved to the same web page. Layout

✕ Bug fixes

  • Fallback to previous webdriver versions of chromedriver, geckodriver, and msedgedriver if the specified browser version's webdrivers are not found.
  • Chromedriver auto download url changed.
  • Probes stats infra chart flickering.
  • Charts for 4-week reports were not displaying correctly.
  • Auto-threshold baselining issue resolved for the 4-week range.

＋ New features

  • Auto-deployment: Web scenarios can now be auto-deployed on less utilized probes. The auto-deployment decision is based on system metrics usage including CPU, memory, and load average. This feature is particularly valuable for load balancing user journey monitoring. Auto-deployment Auto-deployment
  • Auto-signed certificates can be securely exchanged between components in case of IP address migration.
  • All components now come with their own system metrics charts. System metrics System metrics
  • Additional endpoints have been included in the Swagger page of all components. This enhancement further facilitates the creation of custom integrations. Swagger

v1.7.0

This release brings webUI enhancements and tags management. Monitors, URLs and APIs can be grouped by tag groups.

▲ Improvements

  • Cosmetic improvements on the user interface. Layout

＋ New features

  • Tags Management: Monitors, URLs and APIs can be grouped by tag groups. This is particularly useful when you need to group them by specific features and want to display all of them with one click. Plus, now you can toggle on and off the display of tags and tag groups. Tags management Tags management

v1.6.0

This release brings: webUI enhancements (Dashboard, layout). The filtering mechanism based on tags has been enhanced — this further encourages the use of tags to organize one's assets. The color of the tag in the navigation left side bar depicts the worst severity item. Enhancements on user privileges with full and limited profiles. LLD, itemprototype and triggerprototype on Zabbix one-click integration will be automatically created at the template level, with support for Zabbix 6.x. All metrics (including First Contentful Paint, Page Interactive Time, Network Latency, Page Load Time, DNSLookup, TCPConnTime, TLSHandshake, ServerTime) are now sent to influx 1.x and 2.x and can be graphed on Grafana, Splunk or Kibana dashboards. All sensitive data are encrypted in the KV stores.

▲ Improvements

  • Global layout improvements: item (monitors, URLs, APIs) header, colors, etc.
  • Tags filtering improvements: items can be grouped by tags. The color of a tag is based on the worst severity of the tagged items. Colored tags

✕ Bug fixes

  • Tags are correctly defined by users.
  • Filtering toggle not working.

＋ New features

  • Tags filtering. Filtering tags
  • Users can be assigned limited privileges. With full privileges a user can add new tags, while he cannot with limited privileges. User privileges
  • LLD, itemprototype and triggerprototype on Zabbix one-click integration will be automatically created at the template level.
    Support for Zabbix 6.x. Zabbix integration
  • All metrics (including First Contentful Paint, Page Interactive Time, Network Latency, Page Load Time, DNSLookup, TCPConnTime, TLSHandshake, ServerTime) are now sent to influx 1.x and 2.x and can be graphed on Grafana, Splunk or Kibana dashboards. Grafana dashboard
  • Sensitive data are encrypted in the KV stores.
  • Each component comes along with its Swagger page, enabling customized API testing and facilitating seamless integration with other tools. Swagger

v1.5.0

This release brings webUI enhancements. Sparkline graphs are now available for URLs and APIs. Each monitored item has a trend indicator that helps spot which web app is performing better for the past 24h, 7 days and 4 weeks.

▲ Improvements

  • Cosmetic improvements on the user interface. Sparkline graphs are now available for URLs, APIs. Sparkline

＋ New features

  • Trend indicator that helps spot which web app is performing better for the past 24h, 7 days and 4 weeks. Trend Trend
  • Automatic photos will be taken at each start and stop of transaction for monitor items. Photo transaction

v1.4.0

This release brings enhancement on component registration. It fixes bugs on the registration process. Cosmetic enhancements with new icons.

▲ Improvements

  • Cosmetic improvements on the user interface. New icons are being used.

✕ Bug fixes

  • Component registration reads failure while the component has actually correctly self-registered.

＋ New features

  • You can self-register a component from the cli. Self-registration

v1.3.0

This release adds support for monitoring URLs/APIs through basic/authenticated proxies, as well as cosmetic enhancements.

▲ Improvements

  • Cosmetic improvements on the user interface.

＋ New features

  • You can monitor URLs/APIs through basic/authenticated proxies. Similarly the import from CSV file for both URLs and APIs has been updated accordingly. Proxy

v1.2.0

This release brings a new functionality to import a list of URLs from CSV files as well as a list of APIs from CSV files. You start monitoring a bunch of URLs quickly and efficiently.

＋ New features

  • You can import a list of URLs/APIs from CSV files. Import CSV

v1.1.0

This release brings a major new functionality regarding website load testing: the loadtest component, integrated seamlessly with other Mugnsoft core components. Test the performance of your website during peak periods and see how well your transaction can keep up with the load.

＋ New features

  • You can perform your test easily based on your built scenario or on the generated HAR file. In other words you can reuse an existing monitor scenario to build your load test case. Load test
  • You get statistics on the load test plan. From there you can figure out what should be optimized. Load test statistics Load test statistics

v1.0.0

This release brings new charts and new functionalities regarding URL/certificates and API monitoring. The whole process of registering a probe has been refactored. It's now a self pre-registration at the webserver — this brings agility to the process of adding new probes or new integrators. Moreover, the process of exchanging TLS certificates between Mugnsoft's components is now seamless: certificates exchange is automatic. This means that a simple and secure Mugnsoft infrastructure (1 webserver, 1 monitor/probes, 1 integrator) can be deployed in less than 2 minutes.

＋ New features

  • Existing charts can display data from selected probes or locations only. Filter by location or probe
  • New charts — execution count and response time min, max, avg — display data from selected locations only. New charts
  • URL monitoring now includes certificates integrity check and expiration check. URL certificate check
  • API monitoring is now available. You can track the performance of your API endpoints. API monitoring
  • Mugnsoft's components self pre-registration (works seamlessly on Linux):
    Upon installation of a probe it can be self pre-registered. However, its full registration at the webserver should be performed by a webserver admin. Component self pre-registration
  • Mugnsoft's components secure communication:
    Components can communicate securely using either corporate certificates or self-signed ones. Those certificates are then automatically exchanged between trusted Mugnsoft's peers. Secure component communication

v0.1.1

▲ Improvements

  • Network Path displays/hides on HAR switcher select. Network path Network path

✕ Bug fixes

  • Adjustment of the colors of the network path: in green the gateways (IP) which have sent a reply, in red the timeouts and non-responses (which can be quite normal — the provider can simply have set up a rule to filter out ICMP/UDP packets used by the traceroute).

＋ New features

  • Network Path computed for base url of all resources used. Traceroute
  • Ability to emulate Slow Internet Connection (available for Chrome and Edge browsers). Throttling

v0.1.0

Initial release.