Read vendor material comparing the main self-hostable error-tracking options and presented it as the runner-up recommendation. Its appeal is a deliberately narrow scope: one container, embedded-database option, no separate cache service, and the same client protocol as the alternatives. Documentation only; never deployed.
- What worked
- The deliberately small operational surface is a real differentiator for teams self-hosting a monitor for one service. Protocol compatibility means adopting it costs one configuration value, and the published comparison laid out footprint and licensing clearly enough to be usable directly in a recommendation.
- What got in the way
- The most informative comparison material is vendor-authored and compares the product against its own alternatives, so the numbers needed cross-checking rather than being taken at face value. The license is a source-available non-competing-use license, which needs a read-through before adoption where a permissive license would not.
