Tail service events

peren tail reads recent service events from a node. Filter by service or severity to find failed requests and inspect their status, outcome, and wall time. peren logs is the same command under another name.

What is written

The running process appends events to $PEREN_DATA_DIR/tail/events.jsonl (./data/tail/events.jsonl when PEREN_DATA_DIR is unset).

Request events record HTTP dispatch: method, path, status, outcome, and wall time in milliseconds.

Console events record Worker calls to console.debug, console.log, console.info, console.warn, and console.error, with the severity preserved as Debug, Info, Warn, or Error.

Command

peren tail <config> --service <name> [--level log|warn|error] [--node <node>]

--service is required and must name a service in the fleet config. There is no --kind flag. --node is accepted and does not change the result.

The command loads the event file, filters by service and optional level, prints matching lines, and exits. It does not block waiting for new lines.

peren tail fleet.toml --service api

Filter by level

peren tail fleet.toml --service api --level error

Representative output:

api POST /checkout status=500 outcome=error wall_time_ms=37
api console Error checkout failed

What --json is not

peren tail does not accept --json. The printed lines are for you to read. They can change between releases. The file on disk is JSONL. The command does not print that JSON.

Difference from peren logs

None. peren logs and peren tail read the same file, take the same flags, and print the same lines. Pick one name and keep it. See peren logs and the CLI reference.

Failure

Exits 2 if --service is omitted. Exits 1 if the service is absent from config, or if the event log cannot be read after a non-missing I/O error. A missing event file yields no lines and exit 0.