- Maintenance release, updating dependencies.
- Handle slow patchman instances.
- Enable pre-filtering Patchman results, see example.toml for details.
- The menu has been given an overhaul to be more readable in the dark theme.
- The current dashboard is now better recognizable in the menu.
- The 1.10.0 version unfortunately introduced a large unnecessary log volume change.
- Add Graylog event grouping fields to labels.
- Add Graylog connector.
- Only dependency upgrades
- Using
tuwat -otelUrl stdout
orTUWAT_OTEL_URL=stdout
now enables tracing output on stdout. - Add spans/traces to logs to enable correlation
- Cleanup file paths in logging for visibility by stripping the build path.
- Remove error logging keys containing
<nil>
strings.
- Downtimes and acknowledgements as well as disabling notifications for a host
will also disable those for the services on this host. This differs from
icinga
andnagios
, as they allow actions likeSchedule downtime for this host
vs.allow downtime for this host and all services
. - Potentially breaking: Change logging system from
zap
to stdlibslog
. In case the logs are parsed using external systems, the logs are now in a different format. This was done to reduce the overall amount of dependencies.
- Allows times to be in the future (negative seconds ago) and thus to be able to filter those correctly.
- Allow GitLab Groups to be configured for pulling merge requests. This has the advantage of needing to specify fewer single projects and also ignores archived projects by default.
- Fix application crash on certain network timeouts.
- Bugfix release for
when
rules only being applied if nowhere
rule given.
- Bugfix release for
when
rules and~=
matching rule prefixes.
- Add
alertmanager
API, e.g./api/alertmanager/v2/status
for use as API project. Note that this is an experimental feature and should not be relied upon. - Add negative matchers:
!=
and!~
- Use semver for versioning tuwat. This allows tuwat to be used as a library.
- Make the look of tuwat configurable via
style
property.
- The
icinga2
adapter now filters all acknowledged issues.
- Fix rule matching from
v1.1
which made every rule match everything. tuwat
now listens to every network instead of onlylocalhost
for the web port. The management port still only binds tolocalhost
.- The
icinga2
adapter now adds host groups to services. - Add more documentation regarding rules and dashboards.
- Add a
when
rule to be able to match on when the alert happened. - Add more expressive rules, making it possible to express a
greater than
on numerical values. SeeREADME
for details. - Breaking configuration changes:
- The
what
rules are now combined viaAND
with label rules. This streamlines the behaviour, making it behave like the label rules themselves. It also makes it possible e.g. to express that a rule matcher only applies when the alert is old.
- The
- Revise look of alerts
- Add example connector
- Make red alert more readable
- Make filtering Redmine alerts more functional
- Add Redmine connector
- Add
orderview
connector - Recognize service downtimes for
nagios-api
connector - The filtered alert-list is now hidden by default and can be toggled
- UX change: Full row for items now clickable for details
- Adds possibility of having multiple dashboards/rule-sets
- Breaking configuration changes:
- OAuth2 configuration is now a subsection named
OAuth2Creds
- BasicAuth Username now configured with key
Username
instead ofUser
for all connectors.
- OAuth2 configuration is now a subsection named
- Breaking behavioural changes:
- If observability depends on the label
thing
for websocket connections, this has been renamed toclient
.
- If observability depends on the label
- To not display all merge requests, you can now specify the projects via
Projects = []
- Fixes connection leak in nagios-api connector
- Breaking Configuration Changes:
- Change parsing of rules, so that label matchers are combined with
AND
instead ofOR
.
- Change parsing of rules, so that label matchers are combined with
- Handle more open merge requests from GitLab
- Make management port configurable via
-mgmtAddr :8987
- Add
net/pprof
for debugging, see http://127.0.0.1:8987/debug/pprof - Add
Down
state for health endpoint if last collection too old
Breaking Changes:
- actuator endpoints now on different port, no longer on main port (
8988
by default)- http://127.0.0.1:8987/actuator/health (
/info
,/prometheus
)
- http://127.0.0.1:8987/actuator/health (
- Fix websocket/sse registrations
- Make event source selectable via /alerts?eventSource=sse|websocket|fallback
- Use SSE if Websockets are not available
- Add a new GitHub collector showing open issues/pull requests
- Breaking Configuration Changes:
GND_ENVIRONMENT
env var renamed toTUWAT_ENVIRONMENT
GND_INSTANCE
env var renamed toTUWAT_INSTANCE
GND_ADDR
env var renamed toTUWAT_ADDR
-mode dev|prod
removed in favour ofTUWAT_TEMPLATEDIR
andTUWAT_STATICDIR
env vars.
- Added
TUWAT_CONF
env var for specifying the configuration file
- Do not collect, if there is no-one subscribed to the dashboard to avoid cpu/network activity
- Propagate telemetry information to sources
- Open the source under a 3-Clause BSD License
- Preparations for releasing on GitHub
- Rename project from a temporary
gonagdash
to a hopefully less temporarytuwat
. - Revise output on
tuwat -version
- Add preview functionality to silence alerts
- Revise link symbols
- Revise use of GitLab connector, now targeting all visible MRs
- Revise how filters are configured
- More tags for filtering Patchman
- Make collection interval configurable
- Add
SIGHUP
for reloading the configuration
Add links for some connectors and revise the UI to display more information.
Bugfixes all over.
Stream changes to the browser via Websockets.
Initial Release of Tuwat providing a basic non-working Dashboard for Nagios, Icinga 2, Patchman, Alertmanager and GitLab MRs.