In DataStream 2, you can stream logs to third-party destinations for storage, analytics, and enhanced control over your data.
Currently available destinations include Amazon S3, Azure Storage, Datadog, Elasticsearch, Google Cloud Storage, Loggly, New Relic, Oracle Cloud, Splunk, Sumo Logic, S3-compatible destinations, and custom HTTPS endpoints.
Destination features
When choosing a destination to stream logs, you can configure features that differ between available options. This includes custom headers (you can use them for authentication and labeling incoming logs), IP Access List (filtering requests by IP or using Akamaized hostnames as endpoints), dynamic variables (to sort and label logs), and mTLS authentication (for improved security).
While all destinations support the JSON file format, Elasticsearch and New Relic do not support Structured log files. If you're looking for sample logs (JSON and structured) with data, see Log format.
Before choosing the destination to upload your logs, you can consult the table below for features available for each:
| Destination | JSON logs | Structured logs | Custom header | IP Access List | Custom log file prefix & suffix | mTLS authentication | Dynamic variables |
|---|---|---|---|---|---|---|---|
| Amazon S3 | ✓ | ✓ | ✗ | ✗ | ✓ | ✗ | ✓ |
| Azure Storage | ✓ | ✓ | ✓ | ✗ | ✓ | ✗ | ✓ |
| custom HTTPS endpoint | ✓ | ✓ | ✓ | ✓ | ✗ | ✓ | ✗ |
| Datadog | ✓ | ✓ | ✗ | ✓ | ✗ | ✗ | ✗ |
| Elasticsearch | ✓ | ✗ | ✓ | ✓ | ✗ | ✗ | ✗ |
| Google Cloud Storage | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ |
| Loggly | ✓ | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ |
| New Relic | ✓ | ✗ | ✓ | ✓ | ✗ | ✗ | ✗ |
| Oracle Cloud | ✓ | ✓ | ✗ | ✗ | ✓ | ✗ | ✓ |
| S3-compatible destinations | ✓ | ✓ | ✗ | ✗ | ✓ | ✗ | ✓ |
| Splunk | ✓ | ✓ | ✓ | ✓ | ✗ | ✓ | ✗ |
| Sumo Logic | ✓ | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ |
Third-party destination issues
Streaming logs to third-party destinations may temporarily fail due to issues such as latency or connection problems. To prevent data loss during upload failures, DataStream retries data upload up to 10 attempts in 5 minutes for 429 and 5XX error codes. For details, see Delivery retry.
You should take these limitations into account before using data served on your stream for audit, compliance and billing purposes.