OpenLogs
A fast, keyboard-driven viewer for Google Cloud logs that does not copy or store your data
$0product subscription; logs stay in your project
The problem
If your services run on Cloud Run, your logs are already organised and already in Cloud Logging. The tooling around them is the issue. Google's Log Explorer works, but it is slow, cluttered, and built for writing queries rather than for what you actually do when something breaks: filter, search, scan, expand, repeat, fast.
The usual escape is to forward everything to Datadog or BetterStack. That adds a hop, a dependency and a bill, to read data you already own and can already read for free.
What I built
A self-hosted log viewer that reads the Cloud Logging API directly. A thin backend holds the credentials and proxies queries; the frontend is a dense, keyboard-driven reader.
Live tail as entries arrive, severity filters, full-text search across message and metadata, a service switcher for moving between environments, a time-range picker, and structured fields that expand in place. Filter state lives in the URL, so a view you have narrowed down is a link you can paste to someone else.
What it deliberately is not
It is not log storage. Cloud Logging already does retention, and adding a second copy of the data would be the same mistake as forwarding it to a vendor. OpenLogs holds nothing: every view is a query against the API, so there is no index to fall behind, no ingestion to break, and nothing to pay for beyond reads, which are free.
It ships as a single Docker container. Point it at a service account, run it next to whatever else you run, and the logs never leave your own project.
