Pegel
Pegel
Release notes and incidents of Pegel, a small self-hosted uptime monitor. Two people, evenings, one binary. RSS feed
Open incident arm64 downloads of v0.9.0 failed their checksum
Releases
-
v0.9.1: One probe cannot agree with itself
Pegel refuses a config that asks more probes to agree than it has. The release job writes checksums after every build has finished.
pegel checknow refuses a config whoseagreeis larger than the number of probes. In 0.9.0 such a config started without a word and could never alert, because one probe cannot agree with itself. Two people ran into it in the first three days, both on a Raspberry Pi with a single probe.Fixed
- A config with
agree = 2and one[[probe]]started and never fired an alert.pegel checkandpegel runstop and name the line to fix. - Without any
[[probe]],agreedefaults to 1 again, as the notes of 0.9.0 said. pegel proberetried a refused token every second. It waits 30 seconds between attempts now and logs why.
Changed
- The release job writes
SHA256SUMSafter the last build has finished. Details in the checksum incident.
probes := max(1, len(cfg.Probes)) if cfg.Defaults.Agree > probes { return fmt.Errorf("%s line %d: agree = %d needs %d probes, found %d", cfg.Path, cfg.Line("defaults.agree"), cfg.Defaults.Agree, cfg.Defaults.Agree, probes) }$ pegel check --config pegel.toml error: pegel.toml line 3: agree = 2 needs 2 probes, found 1 - A config with
-
v0.9.0: Checks from two places
An alert fires only when two probes agree, which ends the false alarms from a flaky home line. Since this release, interval counts seconds.
Checks can now run from two places at once, and an alert only fires when both see the outage. That ends the false alarms from a flaky home connection, which were most of our bug reports this year. The one breaking change is
interval, which counts seconds now instead of minutes. Pegel refuses to start with an old config and says which line to fix.⚠️intervalcounts seconds, not minutes. Multiply everyintervalinpegel.tomlby 60 before you upgrade, or Pegel will not start.Added
[[probe]]inpegel.tomland the commandpegel probe, which runs your checks on a second machine and reports to your Pegel.agreeunder[defaults]or per check: how many probes must see an outage before an alert fires. It defaults to 2 with two probes or more.- The status page shows which probe saw what.
- Intervals below one minute, down to 10 seconds.
Changed
- Breaking:
intervalis in seconds.interval = 1meant one minute and now means one second, which Pegel refuses. pegel checknames the probes each check ran from.
Removed
- The environment variable
PEGEL_CONFIG, deprecated since 0.7.0. Use--config.
Fixed
- A check that failed and recovered within one interval sent "recovered" without a "down" before it.
Upgrading from 0.8.x
The probe on the second machine gets its checks from your Pegel and needs no config of its own, only the name and a token.
[defaults] -interval = 1 +interval = 60 +agree = 2 + +[[probe]] +name = "home" + +[[probe]] +name = "vps"
The check before the upgrade, on our own config. Why two probes, and what a year of false alarms looked like: Why an alert needs two probes.
-
v0.8.2: Webhook tokens stay out of the log
At log level debug, 0.8.0 wrote webhook URLs with their tokens to the log. Upgrade if you log at debug. There is no v0.8.1.
Pegel no longer writes webhook URLs to its log. At
log_level = "debug", 0.8.0 logged every webhook call with its full URL, and for most chat services the token is part of that URL. If you ran 0.8.0 atdebug, rotate your webhook tokens.Security
- Webhook calls are logged with the host only, such as
hooks.example.com. pegel checkwarns whenpegel.tomlcan be read by other users.
Fixed
- A maintenance window that ended at midnight ended a day late.
-slog.Debug("webhook", "url", hook.URL.String()) +slog.Debug("webhook", "host", hook.URL.Host)There is no v0.8.1. We tagged it on 2026-10-02, its arm64 build failed, and by the time that was fixed the security fix above was ready too. From this release on, a tag builds and publishes in CI only.
- Webhook calls are logged with the host only, such as
-
v0.8.0: Maintenance windows
A check inside a maintenance window still runs but does not alert, and the status page says Maintenance instead of Down.
Planned downtime no longer pages you. A
[[maintenance]]block names the checks and a window. Inside it Pegel keeps checking but sends no alert, and the status page shows Maintenance. Until now the only way was to comment the check out and remember to put it back.[[maintenance]] checks = ["shop", "docs"] start = "2026-08-30T22:00:00+02:00" end = "2026-08-31T01:00:00+02:00"Added
[[maintenance]]withstartandendfor one window.everyfor a window that repeats, such asevery = "sunday 03:00 2h".- Maintenance on the status page, in its own colour.
Changed
- Recovery alerts say how long the check was down.
- Pegel builds with Go 1.27.
Deprecated
quieton a check. A maintenance window does the same and ends on its own.quietgoes in 0.10.0.
Fixed
- The status page counted a paused check as up.
-
v0.7.0: Every check has a name
Checks are named in the config, the alerts and the API. A check without a name stops Pegel from starting. The release also adds pegel check.
Every
[[check]]needs anamenow, and alerts, the status page and the API use it. Until 0.6.0 a check was its position in the file, so moving one check up broke every link to the checks below it.⚠️A[[check]]withoutnamestops Pegel from starting, and API paths change from/api/checks/3to/api/checks/shop. Runpegel checkfrom 0.7.0 against your config before you restart. It lists every check that needs a name.Added
pegel check, which reads the config, runs every check once and exits.- A sound in the dashboard when a check goes down, off by default.
--configto name the config file.
Changed
- Breaking: every
[[check]]needs a uniquename. - Breaking: API paths use the name,
/api/checks/shop. [status]lists checks by name.
Deprecated
PEGEL_CONFIG. Use--config. The variable goes in 0.9.0.
[[check]] +name = "shop" url = "https://shop.example.org/health" expect = 200The dashboard sound for a check that goes down0:00/4.1