bvandeusenandClaude Opus 5.5 4c84bc8560 feat(norms): Steward learns each metric's normal per hour of day, then week (#5739) ...

A metric registered with learn_normal() gets, for every resource reporting
it, a usual hold (the 90th percentile across past days of each hour's p95)
and a usual peak for each slot:
- an hour of the day once 5 days of history exist;
- an hour of the week once history reaches back 3 weeks, so weekly jobs are
  learned too.
Each slot takes in the hours either side, so a job that runs a little late
stays normal. One SQL statement relearns everything hourly into
metric_norms, from the rollup plus the raw window not rolled up yet;
norms_at() reads the slot a moment falls in. Nothing is registered yet:
Traefik (#5741) and hosts (#5742) opt in.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

2026-10-14 12:48:44 -04:00

..

__init__.py

ci: add Forgejo CI (lint + unit + Postgres integration) + ci-requirements.md

2026-06-01 08:37:32 -04:00

_mail.py

feat(email): the report, invites, resets and test sends go through the outbox; only it talks SMTP (#5508)

2026-10-09 16:05:01 -04:00

test_account_time_zone.py

feat(account): change your own password on the Account page (milestone 497)

2026-10-08 20:14:45 -04:00

test_alert_ansible_action.py

feat(ui): each run says who or what started it; run detail without ids (#5212)

2026-10-07 00:26:17 -04:00

test_alert_covering_rules.py

feat(alerts): one rule can cover every host, or every host in a group (#5464)

2026-10-09 08:25:51 -04:00

test_alert_dispatch.py

test(alerts): seed metrics the way ingest does, through record_metrics (#5473)

2026-10-09 09:28:53 -04:00

test_alert_email_outbox.py

test(email): the deleted-alerts outbox test stays in one event loop (#5507)

2026-10-09 16:07:17 -04:00

test_alert_history_page.py

test(alerts): check the history's Action column on one rule's rows, which the test owns (#5316)

2026-10-07 23:06:33 -04:00

test_alert_no_data.py

test(alerts): seed metrics the way ingest does, through record_metrics (#5473)

2026-10-09 09:28:53 -04:00

test_alert_rule_form.py

fix(alerts): a rule's page doesn't wait for its metric preview (#5485); Send test up top (#5486)

2026-10-09 11:44:19 -04:00

test_alert_send_test.py

feat(email): the report, invites, resets and test sends go through the outbox; only it talks SMTP (#5508)

2026-10-09 16:05:01 -04:00

test_alert_states.py

test(alerts): seed metrics the way ingest does, through record_metrics (#5473)

2026-10-09 09:28:53 -04:00

test_ansible_credentials.py

feat(ansible): credentials — SSH key, become, vault, host-key checking (task 548)

2026-06-02 18:41:58 -04:00

test_ansible_foundation.py

feat(ansible): parameterized runs — extra-vars, limit, tags, dry-run (task 546)

2026-06-02 13:31:16 -04:00

test_ansible_inventory.py

fix(ansible): group names Ansible accepts, and interpreter discovery without the warning (#5334)

2026-10-08 00:01:54 -04:00

test_ansible_overlap.py

fix(ansible): overlapping runs: one per host, one per schedule, one action run per alert rule (#5158)

2026-10-06 21:24:51 -04:00

test_ansible_run_end.py

fix(ansible): runs end cleanly: output read in chunks, the process group always ended, a cancelled task recorded (#5156)

2026-10-06 21:08:30 -04:00

test_ansible_runs_list.py

feat(ansible): Inventory on one page, add behind dialogs, delete behind the confirm (#5333)

2026-10-07 23:57:54 -04:00

test_api_tokens.py

feat(auth): sessions end when they should; logins are throttled; Secure behind HTTPS (#5035)

2026-10-06 11:02:27 -04:00

test_backup_job.py

test(backup): run 9076's three failures (#5673, #5674)

2026-10-12 07:27:24 -04:00

test_backup_restore_fixes.py

test(backup): run 9076's three failures (#5673, #5674)

2026-10-12 07:27:24 -04:00

test_boot.py

ci: add Forgejo CI (lint + unit + Postgres integration) + ci-requirements.md

2026-06-01 08:37:32 -04:00

test_cleanup.py

feat(settings): Settings → System shows table sizes and what the last cleanup removed (#5469)

2026-10-09 14:29:55 -04:00

test_dashboard_reset.py

feat(dashboard): a curated default layout, merged duplicate widgets, a labelled Overview (#5213)

2026-10-07 00:42:04 -04:00

test_docker_disk_page.py

test(docker): the disk page's images carry the 12-character ids the agent sends (#5305)

2026-10-07 20:20:41 -04:00

test_docker_recent_usage.py

fix(docker): drop the unused .w-meter style; the bars test reads the loaded plugin's models (#5214)

2026-10-07 15:08:15 -04:00

test_docker_service_page.py

feat(docker): the service and container pages say each fact once (#5310)

2026-10-07 21:33:25 -04:00

test_docker.py

perf(docker): the hourly log prune ranks only the containers over their cap (#5449)

2026-10-08 22:31:00 -04:00

test_email_delivery_pages.py

feat(dashboard): an Alerts widget beside the Overview; the Overview's overflow links to Alerts and Monitors (#5510)

2026-10-09 16:27:17 -04:00

test_email_outbox.py

feat(alerts): alert email waits 2 minutes in the outbox and each alert records its outcome (#5507)

2026-10-09 15:59:17 -04:00

test_external_users.py

feat(auth): LDAP and SSO users keep their source; directory TLS is real (#5036)

2026-10-06 11:10:40 -04:00

test_host_agent_alerts.py

test(alerts): seed metrics the way ingest does, through record_metrics (#5473)

2026-10-09 09:28:53 -04:00

test_host_agent_docker_status.py

feat(docker): Docker's own status reaches Steward — an empty answer clears a host, a lost one greys it, a demoted manager's swarm rows go (#5152)

2026-10-06 18:41:22 -04:00

test_host_metrics.py

perf(metrics): a latest-value table for "what is it now"; dashboard cells say loading, failed and as-of (#5208)

2026-10-06 23:06:34 -04:00

test_host_targeted_run.py

feat(ansible): run a playbook against a single Steward host (task 547)

2026-06-02 19:26:58 -04:00

test_hosts_list.py

fix(hosts): add HX-Request to Vary rather than replacing it; the session adds Cookie (#5294)

2026-10-07 18:57:22 -04:00

test_hourly_p95.py

feat(metrics): each rolled-up hour keeps the level it held, its p95 (#5738)

2026-10-14 12:45:42 -04:00

test_invites.py

feat(email): the report, invites, resets and test sends go through the outbox; only it talks SMTP (#5508)

2026-10-09 16:05:01 -04:00

test_mcp_automation.py

feat(ansible): a run's scope reads as the target or group it names

2026-10-06 10:29:06 -04:00

test_mcp_group_members.py

feat(mcp): update_group_members puts hosts into an inventory group or takes them out (#5487)

2026-10-09 11:48:08 -04:00

test_mcp_monitoring_writes.py

fix(mcp): get_monitor says how often a monitor on the global interval runs (#5433)

2026-10-08 21:50:20 -04:00

test_mcp_plugin_tools.py

feat(mcp): search_container_logs takes a Swarm service's name and searches every task it has run (#5468)

2026-10-09 14:21:56 -04:00

test_mcp_tools.py

perf(metrics): a latest-value table for "what is it now"; dashboard cells say loading, failed and as-of (#5208)

2026-10-06 23:06:34 -04:00

test_mcp.py

feat(mcp): connect card, MCP docs, and an end-to-end roles/audit test over HTTP

2026-10-06 08:25:41 -04:00

test_metric_index.py

perf(metrics): a day of one metric reads from the index alone, built after boot (#5490)

2026-10-09 13:23:46 -04:00

test_metric_latest.py

feat(dashboard): Hosts widget worst first, load against cores, one home per fact (#5251)

2026-10-07 13:15:01 -04:00

test_monitor_history.py

feat(monitors): a monitor's page reads as a page — response time over 24h, labelled actions, how long, not "ago"; the add form says what each type checks (#5214)

2026-10-07 15:57:09 -04:00

test_monitors.py

feat(hosts): the Hosts list labels its figures and says how long; SLA tab gone; one place to add a host (#5279–#5283)

2026-10-07 17:12:17 -04:00

test_norms.py

feat(norms): Steward learns each metric's normal per hour of day, then week (#5739)

2026-10-14 12:48:44 -04:00

test_notification_test_send.py

feat(email): the report, invites, resets and test sends go through the outbox; only it talks SMTP (#5508)

2026-10-09 16:05:01 -04:00

test_passwords.py

feat(email): the report, invites, resets and test sends go through the outbox; only it talks SMTP (#5508)

2026-10-09 16:05:01 -04:00

test_plugin_secrets.py

fix(settings): a plugin's password is sealed at rest, and a blank box keeps it (#5671)

2026-10-12 07:06:31 -04:00

test_prometheus_scrape.py

feat(prometheus): fold state sets on the target page; forget series an endpoint drops (#5689)

2026-10-12 15:11:24 -04:00

test_prometheus_template_library.py

test(prometheus): set the template's version line whatever it is, not only version: 1 (#5693)

2026-10-14 09:45:53 -04:00

test_prometheus_widget.py

feat(prometheus): the Prometheus dashboard widget, target + template (#5691)

2026-10-14 07:05:15 -04:00

test_renames.py

test(alerts): seed metrics the way ingest does, through record_metrics (#5473)

2026-10-09 09:28:53 -04:00

test_schedule_run_now.py

fix(ansible): Run now runs the playbook and leaves the schedule's time alone (#5445)

2026-10-08 22:08:18 -04:00

test_snmp_poll.py

test(snmp): read and remove the poll test's rows in one transaction (milestone 491)

2026-10-08 11:01:08 -04:00

test_traefik_latest.py

feat(traefik): keep each router's 4xx and 5xx responses by status code (#5711)

2026-10-14 11:07:19 -04:00

test_unifi_history.py

fix(unifi): the Overview's history buckets on Postgres; devices name their state (#5207)

2026-10-08 17:28:19 -04:00

test_users_page.py

fix(users): a password reset signs out every session begun with the old password (#5427)

2026-10-08 17:25:35 -04:00