Bitbucket Data Center webhooks to Bamboo Data Center: concurrent push handling and latency

Question

The goal is to determine the documented behavior of Bamboo Data Center when receiving multiple rapid push webhooks from Bitbucket Data Center. Because Bamboo Data Center persists build metadata via Hibernate/JPA to a shared relational database, concurrent plan executions can increase contention on build queue and result summary tables.
While Bitbucket Data Center can deliver push events in rapid succession without guaranteed ordering, Bamboo's trigger handling processes each event independently. There is no documented exactly-once coalescing mechanism. Additionally, the Bamboo REST API utilizes offset-based pagination where query cost increases with offset, potentially contributing to latency spikes that only appear under parallel load.

Does Bamboo Data Center provide any documented deduplication or coalescing logic for rapid successive webhook events targeting the same plan?
  What are the ordering guarantees for builds triggered by concurrent webhooks from Bitbucket?
  How does concurrent database contention affect REST API pagination response times for build and job data?

Tasadduq Burney · Answer

Bamboo Data Center does not provide native deduplication or coalescing logic for rapid successive webhook events. Each push event received from Bitbucket is treated as an independent trigger request. If multiple pushes occur in rapid succession, Bamboo will attempt to initiate a build for every event unless specific plan-level constraints prevent it.

Webhook Handling and Ordering
Because there is no 'exactly-once' coalescing mechanism, the behavior under load is characterized by:
  Bitbucket does not guarantee the order of delivery for concurrent webhooks. Bamboo processes these as they arrive, meaning build execution order may not match the commit history order.
  Concurrent triggers can flood the build queue, potentially exhausting available execution slots.

Database Contention and REST API Performance
Bamboo utilizes Hibernate/JPA to persist build metadata to a shared relational database. Under high concurrent load, the following performance impacts occur:
  Table Contention: Rapid build triggers increase lock contention on the buildqueue and buildsummary tables.
  Pagination Latency: The Bamboo REST API uses offset-based pagination. As the offset increases, the database must scan more rows to reach the result set, leading to latency spikes that are exacerbated when the database is already stressed by concurrent write operations.

Mitigation Strategies
To manage these issues, consider the following steps:
  Webhook Throttling: Implement a middleware layer or a proxy to rate-limit incoming webhooks before they reach the Bamboo API.
  Plan Configuration: Use the "Locking" settings in Bamboo plans to prevent concurrent builds of the same plan, reducing database writes.
  API Optimization: Avoid high offset values in REST API calls; use filtering-based queries to limit result sets and reduce database load.

Bitbucket Data Center webhooks to Bamboo Data Center: concurrent push handling and latency

1 answer

Webhook Handling and Ordering

Database Contention and REST API Performance

Mitigation Strategies

Your perspective belongs here.

0 question comments