How to fix a Spark-808 error within Dynamics / Power Apps?

Question

How to fix a Spark-808 error within Dynamics / Power Apps?

Megan Coldeweijer 0

We have set up a Synapse Link connection within Power Apps. However since we performed a Spark Pool update (from 3.4 to 3.5) a couple of tables show a Spark-808 error: Spark compute pool in Synapse workspace is insufficient. One or more tables could not be updated. About 20 tables are affected with this error. We increased number of nodes by one in the configuration of the (new) Spark Pool, but this didn't fix this error. What else can we do?

0 comments

1 answer

Your answer

Answer 1

Smaran Thoomu 35,870 Microsoft External Staff Moderator

Hi @Megan Coldeweijer
Thanks for the details - increasing nodes alone doesn’t always resolve this error.

After a Spark runtime upgrade (3.4 → 3.5), we do sometimes see Spark-808 when the workload becomes more resource-intensive or when multiple tables are processed at the same time.

A few things I’d suggest checking:

Try running sync for a smaller set of tables (or even one of the impacted tables) to see if it completes — this helps confirm if it’s resource contention
Check if the Spark pool is getting queued jobs or high utilization during sync
Instead of only increasing node count, try using a larger node size (more memory), especially if those tables are bigger
Restart the Spark pool once and retry - sometimes needed after runtime upgrade
If possible, re-trigger sync for one affected table to validate

Since ~20 tables are impacted, this does point more towards capacity/concurrency on the Spark pool rather than a configuration issue in Power Apps.

If you can share your current Spark pool size and whether autoscale is enabled, I can help you fine-tune it further.

Megan Coldeweijer 0 Reputation points

2026-04-03T06:07:13.3366667+00:00

Hello Smaran,

Thank you for your response. I will share your suggestions with my colleague and we will get back to you!
Megan Coldeweijer 0 Reputation points

2026-04-07T10:37:02.7566667+00:00

Hello Smaran,

We have increased the memory and that didn't work. We tried switching the tables off and on again within Synapse Link, but the status immediately gave the 808-Spark error upon enabling it back.

We think that why the error remains is maybe related to the pay-as-you-go subscription since that has limitations, but we are not fully sure.

We would have liked to restart the Spark Pool to see if that worked, but I didn't noticed an option for that. Just an option to pause synchronisation (temporarily).

Do you have any other tips maybe?
Smaran Thoomu 35,870 Reputation points Microsoft External Staff Moderator

2026-04-07T15:45:28.97+00:00
Hi Megan Coldeweijer

Thanks for the update - this helps narrow it down.

Since increasing memory didn’t help and the error appears immediately after enabling tables, this is likely not just a sizing issue. Also, your assumption about pay-as-you-go is understandable, but this error is not typically caused by subscription type, it’s more about how Synapse Link jobs are being handled in the Spark pool.

A few things you can try next:

Instead of enabling many tables together, try enabling just 1 affected table and see if it still fails immediately. This helps confirm if it’s not concurrency-related.

Try stopping (pausing) the Spark pool and starting it again - in Synapse this is effectively the “restart” option.

Check if there are any stuck or queued Spark jobs in the monitoring tab - sometimes jobs from previous runs block new executions.

Since this started after moving to 3.5, try creating a new Spark pool (fresh) with the same config and test Synapse Link against it - this helps rule out any backend/runtime state issue.

If possible, reduce parallelism by enabling tables in smaller batches rather than all at once.

If the error still occurs immediately even for a single table on a fresh pool, we may need to look at this from a backend perspective.

If you can share your Spark pool config (node size, autoscale settings), I can suggest more targeted tuning.