Regex to parse deployment or job name from a Kubernetes pod name

Question

I am trying to parse Kubernetes Pod names for logging.

My pod names always look like this

<deployment>-<replicaset>-<uid>
<job>-<uid> <-- if created by a job

Here are some samples

events-worker-7c9b7bdc55-f7sgc
notification-585f6b94b8-t4jjc
report-generator-749ccf648d-gd9j7
static-content-8445d7f556-wbxvp
init-database-fm44h <-- if created by a job

What I am trying to get is the <deployment/job> part. For above samples this would be

events-worker
notification
report-generator
static-content
init-database

I started with something like this

(?<role_name>.*)(?:-[a-z0-9]{8,10})(?:-[a-z0-9]+)

and ended with this

(?:(?<role_name>[a-z0-9]+(?:-[a-z0-9]+)*))-(?<=-)[a-z0-9]+-(?:(?<=-)[a-z0-9]+)

But I am unable to match both cases (when the name has a replicaset and if it has none)

It either does not match init-database-fm44h at all or only captures init instead of init-database.

Any help would be greatly appreciated

Answer 1

You can use

\b(?<!-)(?<role_name>[a-z0-9]+(?:-[a-z0-9]+)*?)(?:-([a-f0-9]{10}))?-([a-z0-9]+)\b(?!-)

See the regex demo.

Details:

• \b(?<!-) - a word boundary not immediately preceded with -
• (?<role_name>[a-z0-9]+(?:-[a-z0-9]+)*?) - Group "role_name" with ID 1: one or more letters or digits and then zero or more sequences of - and one or more letters/digits as few times as possible
• (?:-([a-f0-9]{10}))? - an optional non-capturing group matching a - and then ten hex chars captured into Group 2
• - - a hyphen
• ([a-z0-9]+) - Group 3: one or more letters or digits
• \b(?!-) - a word boundary not immediately followed with -.