Releases 14.0.15
Compare
Sorry, something went wrong.
No results found
Changes
Notes from an event's overview use its Notes page composer: one step, with templates, AI and who it reaches
Notification methods: say when SMS, calls, WhatsApp or Telegram are off, with the switch right there
Reminders, alert privacy, status page visibility, runbook, SLO evaluation and session replay switches save when flipped
Form guards' "are really read" floors sit below what simplification can reach (CI runs without ee/)
One Probes & Interval page for a monitor's probes, interval and probe agreement
Forms of three rows or fewer are one page (note and postmortem templates, OID templates, SLO create), and a guard keeps it that way
Dashboard custom domains keep a working certificate: verified, ordered and renewed by workers, never deleted by the status page job
Project Settings and Admin one-switch settings save on flip, and ask first where they can lock people out
test(e2e): judge an explorer double-click's landing only once its timing held
SSO: adding an OIDC provider asks for the issuer, client ID and secret; the rest is filled in
test(e2e): the chart zoom fixture's analytics lists answer like the server
fix(database): stop a project create from taking the tenant project's id
Fix master CI: ee SSO license tests (#4319 , #4322 ) and the SLO e2e Description fill (#4318 )
AI settings: each AI behaviour is a switch that saves on flip, limits folded under Advanced
ci(release): tag each release on the commit the run built
SSO: adding a SAML provider or a SCIM connection asks only what the identity provider gives you
Create Team asks the team's Access (as API keys do); Block Permissions fold under Advanced on a team's Permissions page
Admin Dashboard: adding someone to a project starts on its Members team, in one step
Status page custom domains: add it, the DNS record opens right away, and the free certificate is issued without a button
Admin LLM, Call & SMS, WhatsApp, Invite User and calendar link settings lose steps they didn't need; the Advanced-step guard reads translated titles
Status pages: who can see a page is one choice on Access (anyone, people who sign in, or anyone with the password)
No always-on banners on session replay, inventory and database pages; replay retention edited in one place
Share a dashboard with one 'Who can view' choice, from the dashboard's ⋯ menu
Custom domain certificates: one Let's Encrypt budget, one order per name, nothing deleted on a DNS blip, and failed orders shown
Each resource's name, description, labels and identity are edited in one place: its Settings page
Incoming call escalation rules: one step - who to call and how long to ring
Dashboard custom domains work like status page ones: DNS Setup on create, Check now orders the certificate, no Order Free SSL
Create an announcement in two steps, with the status page already picked
A project's first Twilio config becomes its default, so SMS and calls use it at once
Scheduling maintenance and announcing suggest the status pages that show the affected monitors, one click to add
On-call rules: one page with a tab per kind, in User Settings and under Users
Status pages: overall uptime % and what counts as downtime are rows of 'What your status page shows', each saving on its own
Log and trace data rules start from what the rule does; scrub rules that would scrub nothing are refused
chore: npm audit fix
Creating from a record's tab keeps that record picked: Declare Incident, Create Alert and maintenance from monitors, hosts, clusters, services and more
fix(telemetry): store a NATS broker's split delivery from OBI as SERVER
fix(traces): include the spans under a span in its Profile tab
fix(telemetry): drop OBI v0.14's own Node.js inspector requests
feat(kubernetes-agent): ask OBI for span metrics as application_span_otel
Stepped forms: the primary action only on the last step, and Next is never primary
Forms and pages fold rarely used options under More fields / More settings, and the folded header shows what is inside
Declare Incident: monitors apart, Change Monitor Status to only with a monitor, status page limit and notify under More fields
Subscribers who picked a monitor group hear about announcements and maintenance notes on its monitors
Status page: subscribing is one page again, with Preferences folded to what the visitor will get
test(kubernetes-agent): run the agent's eBPF pipeline end to end on a real kernel
fix: write NUL characters as \u0000 so git diffs these files as text
Dropdown and filter-chip clear buttons sit beside their trigger, not inside it; no control nested in another, with a guard
Kubernetes Control Plane and Service Mesh: the Helm setup hint shows only when a tab's charts are empty
fix: master CI, js-lint's 3 errors and Compile App's heap out of memory
Incidents, alerts and maintenance events show their own measurements on their pages
Postmortem: Apply Template only with a template, publish options in order, and settings paths that exist
Card descriptions say what each list or card is for, not 'Here is a list of…', with a guard
fix(telemetry): store an MQTT or NATS broker's client-typed publish as SERVER
fix(profiles): count only samples on a trace's own spans in its Profile tab
Products menu: Essentials are always expanded and can no longer be folded
Card action buttons such as Edit sit on the right of the card header, everywhere
Created sits on the ID line instead of a row of its own, on every details card
Search (Cmd+K) finds every settings page and action, such as API Keys or Delete Project, and searches better
Status page email reports: one switch with a default schedule, and turning them off needs no dates
feat(forms): Branding section - a form's own logo, alt text and favicon
Dashboards: picking a template fills in the name, and Create opens the new dashboard
One incident role picker everywhere, no 'Multiple' tag, and incident templates start with no state
New labels, states, severities, statuses and incident roles start with a colour already picked
Grouping rules make people owners of the episodes they open, replacing a Default Assign To nothing showed
Scheduled maintenance: monitors picked apart from other affected resources, Change Monitor Status to right under them
Log, trace, exception and security event monitors fold their extra filters under More fields
Metric pipeline rules and recording rules start from the rule: no Basic Info step, All/Any only with two filters
Label and owner rules on 21 products open on what they match, then what they add (no Basic Info step)
You can’t perform that action at this time.
