The influxdata-archive.key is designed to have a good user experience
during key rotations, but in order to do that, we need to properly
verify the key file's fingerprint and not its sha256sum (since the
primary key's fingerprint won't change with key rotations, but the
sha256sum necessarily will).
Update documentation for this to establish best practice and prepare
users for the upcoming key rotation.
Adjust dockerfiles to use this methodology.
- Adds basic support for core and enterprise in getswagger.sh
- Adds custom OpenAPI info for Core and Enterprise
- Validates as OpenAPI 3.0 (using Spectral)
- operationId
- tags
- Revises use of Legacy, v2
- TODO: need to check and validate in UI, adjust tags if nec.
- Add and remove components
- Update parameters
- Add examples
- Add tests for Core
- Updates Linux install instructions to reflect updates in the Downloads page.
- Uses curl, instead of wget, for examples.
- Reorg of download verification: moves expandable details before instructions, to encourage using verification.
Includes verification in instructions.
- All tests pass if not explicitly skipped (for Docker and different architectures).
- Run `docker compose build` to update the Go version and other dependencies used by the
test image.
- Updates Telegraf install instructions, adds more info to promote verifying downloads.
- Adds setup for tests.
- Makes tests pass in install and configuration docs.
- Adds Vale.sh exception to allow duration values inside code blocks.
- Adds Dockerfiles for convenience to test install procedures in CentOS and Ubuntu.
- Clarifies and fixes config file handling and specifying input and output plugins.
- Fix the http-write-timeout definition, which should be similar to https://pkg.go.dev/net/http#Server.WriteTimeout
- Add a Vale spelling config for v2 server options
- Add influxdb to the test container packages (for testing service influxdb...)
- Add Dockerfile config and test setup for testing some InfluxDB startup config options (using influxd)
- Pre-commit hook:
- Redo test harness so commands can run in parallel
- Cleanup Vale command args
- Replace lint-staged and husky with Lefthook.
- Pre-push hook:
- Scan for package vulnerabilities
- Tests
- Run tests on demand, in parallel, or debugged using docker compose up, run, exec, etc.
- Each test container (service) runs the prepare-content script before running tests.
- Return to using Python instead of Docker to provide environment variables so they don't get exposed in Docker inspect.
- Bind mount influxctl config.toml and .env.test files
- Bind mount scripts for faster builds while iterating on test scripts.
- Vale:
- Override Acronyms style def--ignore acronym if it's linked.
- For influxctl OAuth flow, add support for writing auth URLs to a container-shared file monitored by the host during the pre-commit hook.
- In prepare-content.sh, add Management API-associated placeholder substitutions.
- Update CONTRIBUTING.md with env.test requirements.
- Reconfigures prettier linting.
- Adds .editorconfig to help with consistent editor settings
- Refactors test runs:
- Removes test configuration from compose.yaml (not suited for this use case).
- Splits test runner into test content setup and pytest that can be run separately or together (and with other test runners in the future).
- Configuration is in Dockerfiles and command line (`.lintstagedrc.mjs`)
- Updates CONTRIBUTING.md
- Updates client library write examples in cloud-dedicated and clustered.
Jamie Strandboge
Jason Stirnaman