Shard ID fixes
We noticed a problem about some shard IDs failing to calculate using free APIs (polygon-rpc.com) and optimized the assignment list loading using our go-multicall library. This also helps reducing the load on the registry APIs and increases the general consistency.
Another problem that is fixed along with this release is the mixup in the shard ID that appears in the metrics. This helps avoid confusion while monitoring how the network operates bot shards.
New forta status all
subcommand
Now instead of using forta status --show all --format oneline
, operators can simply use forta status all
!
More metrics
We added more metrics to increase visibility in the issues:
- Errors during loading assigned bots
- Bot alert processing response validation errors
Reduced docker pull
concurrency
To remedy the pull retry issues visible in some nodes, we have reduced the concurrency to pull one image at a time and increased the timeout from 5 minutes to 10 minutes. This will help speed up the eventual start of bots.
Local mode improvements
- Fixed standalone mode which was broken starting with v0.8.0
- Fixed the zero block error which happens when working with Hardhat
- Disabled the trace requirement for all chains
Other improvements
- Fixed bot attachment bug that happens during graceful auto-upgrades
- Started the JSON-RPC API health checker
- Pinned
docker/docker
(moby/moby
) version to the minimum requirement v20.10.x after dependency update issues - Reduced test coverage fluctuation between test runs (due to concurrent parts)
What's Changed
- Disable trace requirement in local mode by @canercidam in #749
- Reduce the heat on docker image pulls by @canercidam in #756
- Start the JSON-RPC proxy API health checker by @canercidam in #754
- Produce alert response validation error metrics by @canercidam in #753
- Fix shard ID mixup in metric details by @canercidam in #751
- Make bot launch work with running bots after service upgrade by @canercidam in #750
- Add new 'forta status' shorthand 'forta status all' by @canercidam in #755
- Fix standalone mode by @canercidam in #752
- Fix docker pull worker concurrency by @canercidam in #758
- Add error metrics for loading assigned bots by @canercidam in #759
- Pin Docker version to v20.10.x latest by @canercidam in #763
- Fix Hardhat zero block panic by @canercidam in #762
- Migrate registry store to use multicalls by @aomerk in #761
- Fix multicall problem in e2e by @canercidam in #764
- Reduce fluctuation in coverage by @canercidam in #765
- Stabilize test coverage further by @canercidam in #766
Full Changelog: v0.8.3...v0.8.4