Redis
Redis Sink
Overview
The Redis sink publishes observability data to a Redis instance and can be used as a lightweight buffering, queuing, or fan-out mechanism within a telemetry pipeline. It supports logs, metrics, and traces, making it suitable for both transient streaming use cases and short-term data staging.
This sink is stateless and optimized for fast ingestion, with configurable buffering, encoding, acknowledgement handling, and support for multiple Redis data structures.
Redis Data Types
list – Queue-like structure using push/pop semantics
channel – Pub/Sub based broadcasting to multiple subscribers
sortedset – Score-based ordered storage, suitable for time or priority ordering
Required Configuration
inputs List of upstream source or transform IDs.
endpoint Redis connection URL in the form of protocol://host:port/db. Both redis and rediss (TLS-enabled) schemes are supported.
key Redis key used to publish messages. Supports dynamic templating based on event fields.
encoding.codec Defines how events are serialized before being written to Redis.
Encoding
The Redis sink supports multiple encoding formats, including:
json
text
raw_message
logfmt
csv
cef
gelf
protobuf
otlp
avro
Encoding configuration determines which event types (logs, metrics, traces) are supported and how data is serialized before being sent to Redis.
Buffering
Optional buffering improves throughput and resilience during downstream slowdowns:
memory – High performance, volatile
disk – Durable, survives restarts
Behavior when full:
block (default)
drop_newest
Acknowledgements
The Redis sink supports sink-level end-to-end acknowledgements. When enabled, upstream sources that support acknowledgements will wait until events are successfully written to Redis before marking them as delivered.
List Options
For list data type, the following push methods are supported:
rpush – Append messages to the tail of the list (default)
lpush – Insert messages at the head of the list
Sorted Set Options
For sorted set data type, messages are inserted with an associated score. Scores can be static or dynamically derived from event fields using templates.
TLS and Security
Secure Redis connections are supported using the rediss scheme or explicit TLS configuration:
Encrypted connections
Optional certificate verification
Support for secure and insecure TLS modes
Redis Sentinel Support
The Redis sink can connect via Redis Sentinel for high availability deployments:
Sentinel service discovery
Configurable authentication
RESP2 and RESP3 protocol support