What a Unix timestamp is
A Unix timestamp counts the seconds since 00:00:00 UTC on 1 January 1970, ignoring leap seconds. It has no time zone and no daylight-saving rules, which is exactly why logs, databases and alert payloads use it: two systems that agree on the number agree on the moment.
Seconds, milliseconds or nanoseconds?
- 10 digitsSeconds. What
date +%s, most Unix tools and JWTexpclaims use. - 13 digitsMilliseconds. JavaScript's
Date.now(), Java'scurrentTimeMillis()and many JSON APIs. - 16 or 19 digitsMicroseconds or nanoseconds. Go's
UnixNano(), Prometheus internals and some tracing systems.
The tool guesses the unit from the size of the number. A mismatch is the classic bug: a seconds value fed to a milliseconds API lands in January 1970, and a milliseconds value fed to a seconds API lands tens of thousands of years in the future.
A signed 32-bit integer runs out at 2,147,483,647, which is 03:14:07 UTC on 19 January 2038. Systems that still store timestamps that way wrap to December 1901. Modern platforms use 64-bit values, but embedded devices, old file formats and a surprising amount of C code do not.
Reading incident timelines
When you reconstruct an outage, convert every source to UTC first: the load balancer log, the deploy webhook, the first alert, the first customer ticket. Mixed zones are the most common reason a post-incident timeline does not add up. SutramX records every check, alert and incident in UTC and shows them in your zone, and the cron parser uses the same convention for scheduled jobs.