I’m facing an issue where Bitbucket Pipelines cannot reliably connect to my AWS RDS instance when using 1× or 2× runner sizes, even though my RDS security group contains all official Atlassian outgoing IPv4 ranges.
The same pipeline works 100% of the time when using:
options: size: 4x cloud: true
This strongly indicates that smaller/local runners may be using non-documented or dynamic egress IPs, causing random connection timeouts.
Could you please confirm:
Do 1×/2× runners use additional or dynamic IPs not listed in the official documentation?
Is there any guarantee that non-cloud runners use only the documented IP ranges?
Is using cloud: true the only supported method to get static egress IPs?
- what shoild i do if i want to use 2x or default 1x with our integration tests on private RDS
