feat: add new RDS metrics to eporter

[deeb00]

Hey everyone! 👋

First of all, thanks for this awesome exporter - it's been super helpful in monitoring our infrastructure!

While using it, I noticed we were missing some important RDS metrics that could give us better insights into our database performance. So I decided to add them! Here's what's new:

CPU & Credit Metrics:

• BurstBalance - to track gp2 burst-bucket I/O credits
• CPUCreditBalance & CPUCreditUsage - for monitoring CPU credits
• CPUSurplusCreditBalance & CPUSurplusCreditsCharged - for tracking surplus CPU credits

EBS & Storage Metrics:

• EBSByteBalance & EBSIOBalance - for monitoring EBS volume performance
• DiskQueueDepth - to track outstanding I/O operations
• ReadLatency & WriteLatency - for measuring disk I/O performance

Network & Replication:

• NetworkReceiveThroughput & NetworkTransmitThroughput - for network traffic monitoring
• OldestReplicationSlotLag - to track replication delays
• CheckpointLag - for monitoring WAL consistency
• TransactionLogsGeneration - to track transaction log activity

These additions should give us a more complete picture of our RDS instances' performance and resource utilization. Let me know if you'd like any adjustments or have questions!

Cheers! 🚀

[serge-r]

Good job, we have to use this metrics too, using YACE for it. Would be nice to have full set of metrics in one exporter.

[deeb00]

Hey! 👋

I did another round of testing after submitting the PR and noticed that the metrics weren't being collected. Here's what I saw in the debug output:
{"time":"2025-04-21T13:54:00.978701+07:00","level":"DEBUG","msg":"cloudwatch metrics fetched","metrics":{"Instances":null}}

Turns out CloudWatch has a special requirement when it comes to query IDs - they can't contain special characters like %. But we need to keep the % in the actual metric names when making API calls.

I've fixed this by:

• Stripping % from query IDs
• Keeping the original metric names with % for API calls
• Making sure the responses are processed correctly

Now all the EBS balance metrics (EBSByteBalance% and EBSIOBalance%) should be collected properly!