Timestream write_records is orders of magnitude slower over the last 3 days in eu-west-1


I have a python lambda streaming data to timestream that has been running for the last few months.
The lambda has always run well under a second, with an average of 15-30ms - regardless if the data itself is current (memory store) or historical (magnetic store).
This week the lambda is timing out, even when the timeout duration is set to a max of 15 mins.
The place its getting stuck is where we make the call to timestream to write the actual records.
It just sits here, for a duration of between 30ms to 15mins.
The data that gets stuck is always processed on a retry and appears in the DB, so its not a data issue.
This is causing our SQS queues to build up, as our lambda failures and timeouts cause throttling, and a reduced concurrency.

The line in question :
result = write_client.write_records(DatabaseName=DATABASE_NAME, TableName=table_name, CommonAttributes=common_attributes, Records=records)

asked 3 months ago35 views
1 Answer

hello. just to add this is still happening - 40% of the time it executes a write in ~500ms other times 5secs roughly 40% of the lambda times out - up to 15 mins.

answered 3 months ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions