🗄️ Archiver

Task for archiving old RapidPro/TextIt runs and messages. It interacts directly with the database and writes archive files to S3.

It runs a single archival pass and then exits. The exit code is meaningful: 0 on success, non-zero if any initialization step or the archival itself failed. It does not schedule itself — run it on whatever cadence you need from an external scheduler (cron, a Kubernetes CronJob, an AWS ECS Scheduled Task driven by EventBridge, etc).

Configuration

• ARCHIVER_DB: URL describing how to connect to the database
• ARCHIVER_VALKEY: URL describing how to connect to Valkey

AWS

The AWS region and credentials are resolved via the standard AWS SDK default chain — the AWS_REGION (or AWS_DEFAULT_REGION) environment variable, the instance/task IAM role, AWS_ACCESS_KEY_ID / AWS_SECRET_ACCESS_KEY environment variables, the shared config/credentials files, etc. When running on AWS the task/instance role is the recommended option for credentials.

For writing of archives, Archiver needs access to a storage bucket on an S3 compatible service. We recommend that you choose SSE-S3 encryption as this is the only type that supports validation of upload ETags.

• ARCHIVER_S3_BUCKET: name of your S3 bucket (e.g. dl-archiver-test")

If using a different encryption type or service that produces non-MD5 ETags:

• ARCHIVER_CHECK_S3_HASHES: can be set to FALSE to disable checking of upload hashes.

Logging and error reporting:

• ARCHIVER_DEPLOYMENT_ID: used for metrics reporting
• ARCHIVER_SENTRY_DSN: DSN to use when logging errors to Sentry
• ARCHIVER_LOG_LEVEL: logging level to use