Skip to content
Akema Docs

Akema documentation

Akema is a read-only lens on Apache Kafka. It reads the topics you choose into memory, shows the messages in a live table, tags the ones that matter with CEL traps, and charts a window of traffic with DuckDB in your browser. It never writes to Kafka, and it never stores message content on disk.

Start here

Section titled “Start here”
  • Install Akema on Linux, Docker, macOS or Windows, and start the 14-day trial.
  • Quickstart: load a sample file, trap failed payments with CEL and chart them in five minutes, no Kafka needed.
  • Tutorial: run Kafka in Docker, stream a small shop into it, and catch a payment outage and lost payments in about 30 minutes.
  • Connect to Kafka: settings for self-managed Kafka, Amazon MSK, Confluent Cloud and Azure Event Hubs.
  • Traps and CEL: tag and count the messages that matter, with a CEL reference and recipes.
  • Explore and charts: build charts by dragging fields, filter by clicking, and present or query the results.
  • Settings: control what Akema keeps, how it streams, and how it copes with busy topics.
  • PII redaction: pseudonymize personal data before Akema shows, traps or charts it.
  • Sign-in and sharing: share one Akema with your team, with one editor at a time and an optional sign-in with editor and viewer roles.
  • Serve over HTTPS: optionally encrypt the connection to Akema’s page, with a local certificate Akema makes or one you already have.

What Akema works with

Section titled “What Akema works with”
  • Self-managed Apache Kafka (including KRaft), Amazon MSK, the Azure Event Hubs Kafka endpoint and Confluent Cloud
  • TLS, mutual TLS, and SASL with PLAIN, SCRAM-SHA-256 or SCRAM-SHA-512
  • JSON, Avro and Protobuf values, through Schema Registry or schema files you upload
  • CSV, JSON and Parquet files, so you can explore without a cluster

Use cases

Section titled “Use cases”
  • Log aggregation: catch error bursts, slow requests and sensitive audit events in application logs.
  • Event-driven order pipeline: catch failed payments and prove no order skipped a step.
  • Clickstream: catch bots that buy and searches that find nothing, and see revenue by campaign.
  • Change data capture (CDC): catch deletes and oversold stock, and find changes that would roll a row back.
  • IoT and metrics: catch overheating machines from raw readings and check that no alert was missed.
  • Telecom usage records: catch duplicate and unknown-subscriber records and check mediation before billing.

Something not working? See Troubleshooting. Wondering whether Akema fits? See Limitations. Questions this documentation doesn’t answer yet? Email sales@akema.io.