panic: runtime error: invalid memory address or nil pointer dereference

[ziansong]

I am running 0.8.0 app version with aws role that has admin access When I run this app locally against Account A (with only mysql db instance) it worked However when I run this app locally against Account B (this one has multiple instances with mysql postgreSQL) I get the below error

{"time":"2024-03-18T03:03:37.915452047Z","level":"INFO","msg":"starting the HTTP server component"} {"time":"2024-03-18T03:03:52.446299831Z","level":"INFO","msg":"get RDS metrics"} panic: runtime error: invalid memory address or nil pointer dereference [signal SIGSEGV: segmentation violation code=0x1 addr=0x20 pc=0x1c11be2]

goroutine 100 [running]: github.com/qonto/prometheus-rds-exporter/internal/app/ec2.(EC2Fetcher).GetDBInstanceTypeInformation(0xc00046df40, {0xc000434400, 0xc, 0x10}) /home/runner/work/prometheus-rds-exporter/prometheus-rds-exporter/internal/app/ec2/ec2.go:94 +0x842 github.com/qonto/prometheus-rds-exporter/internal/app/exporter.(rdsCollector).getEC2Metrics(0xc00015c508, {0x7f700a6438c0, 0xc000292680}, {0xc000434400, 0xc, 0x10}) /home/runner/work/prometheus-rds-exporter/prometheus-rds-exporter/internal/app/exporter/exporter.go:436 +0x14f created by github.com/qonto/prometheus-rds-exporter/internal/app/exporter.(*rdsCollector).fetchMetrics in goroutine 97 /home/runner/work/prometheus-rds-exporter/prometheus-rds-exporter/internal/app/exporter/exporter.go:376 +0x537

[ziansong]

ADD： I only turn on PROMETHEUS_RDS_EXPORTER_COLLECT_INSTANCE_TYPES: true turn off other If delete ec2 privilege in aws iam role I can get rds_instance_info metrics

[Antoine-Sevec]

I have the same problem, do you have a solution? Thanks

[vmercierfr]

Thanks for reporting this issue, sorry about that issue. It's definitely a bug that we must fix!

I'll propose fix by the middle of next week.

[vmercierfr]

We just released the 0.8.1 that contains the fix.

Can you update to this version and confirm that it works as expected?

[Antoine-Sevec]

Bonjour, merci pour le retour.

Je viens de tester votre nouvelle version 0.8.1 et j'ai comme erreur :

│ {"time":"2024-03-25T15:24:36.471717476Z","level":"INFO","msg":"starting the HTTP server component"}                                                                                                                                                                                                                        │
│ {"time":"2024-03-25T15:24:47.29626138Z","level":"INFO","msg":"get RDS metrics"}                                                                                                                                                                                                                                            │
│ panic: runtime error: invalid memory address or nil pointer dereference                                                                                                                                                                                                                                                    │
│ [signal SIGSEGV: segmentation violation code=0x1 addr=0x20 pc=0x1c206ff]                                                                                                                                                                                                                                                   │
│                                                                                                                                                                                                                                                                                                                            │
│ goroutine 111 [running]:                                                                                                                                                                                                                                                                                                   │
│ github.com/qonto/prometheus-rds-exporter/internal/app/ec2.(*EC2Fetcher).GetDBInstanceTypeInformation(0xc0006b9f40, {0xc000540a00, 0x9, 0x10})                                                                                                                                                                              │
│     /home/runner/work/prometheus-rds-exporter/prometheus-rds-exporter/internal/app/ec2/ec2.go:95 +0x8bf                                                                                                                                                                                                                    │
│ github.com/qonto/prometheus-rds-exporter/internal/app/exporter.(*rdsCollector).getEC2Metrics(0xc0000f8a08, {0x7fe2caf7d440, 0xc0002fa4e0}, {0xc000540a00, 0x9, 0x10})                                                                                                                                                      │
│     /home/runner/work/prometheus-rds-exporter/prometheus-rds-exporter/internal/app/exporter/exporter.go:436 +0x14f                                                                                                                                                                                                         │
│ created by github.com/qonto/prometheus-rds-exporter/internal/app/exporter.(*rdsCollector).fetchMetrics in goroutine 62                                                                                                                                                                                                     │
│     /home/runner/work/prometheus-rds-exporter/prometheus-rds-exporter/internal/app/exporter/exporter.go:376 +0x537

Voici ma configuration values.yaml :

#
# Exporter settings
#

# Enable debug mode
debug: true

# Log format (text or json)
# log-format: json

# Path under which to expose metrics
# metrics-path: /metrics

# Address to listen on for web interface
# listen-address: ":9043"

# Path to TLS certificate
# tls-cert-path: ""

# Path to private key for TLS
# tls-key-path: ""

# Enable OpenTelemetry traces
# See https://opentelemetry.io/docs/languages/sdk-configuration/otlp-exporter for configuration parameters
# enable-otel-traces: true

#
# AWS credentials
#

# AWS IAM ARN role to assume to fetch metrics
# aws-assume-role-arn: arn:aws:iam::000000000000:role/prometheus-rds-exporter

# AWS assume role session name
# aws-assume-role-session: prometheus-rds-exporter

#
# Metrics
#

# Collect AWS instances metrics (AWS Cloudwatch API)
collect-instance-metrics: true

# Collect AWS instance tags (AWS RDS API)
collect-instance-tags: true

# Collect AWS instance types information (AWS EC2 API)
collect-instance-types: true

# Collect AWS instances logs size (AWS RDS API)
collect-logs-size: true

# Collect AWS instances maintenances (AWS RDS API)
collect-maintenances: true

# Collect AWS RDS quotas (AWS quotas API)
collect-quotas: true

# Collect AWS RDS usages (AWS Cloudwatch API)
collect-usages: true

[ziansong]

I have the same problem, do you have a solution? I only use collect-instance-types: true

[vmercierfr]

Sorry about that! I'm reopening this bug and will propose a more complete fix.

[vmercierfr]

The AWS SDK could return a null pointer for RDS instances that do not support optimized EBS, we were not testing this case and https://github.com/qonto/prometheus-rds-exporter/pull/152 was not enough to handle it.

We adopted a more robust approach with https://github.com/qonto/prometheus-rds-exporter/pull/154.

We have just released the v0.8.2, which should finally solve this problem.

@Antoine-Sevec @ziansong Can you update to this version and confirm that it works as expected?

[Antoine-Sevec]

@vmercierfr

C'est résolu pour moi, merci pour la réactivité !

│ {"time":"2024-03-26T12:03:35.208829158Z","level":"INFO","msg":"starting the HTTP server component"}                                                                                                                                        │
│ {"time":"2024-03-26T12:03:47.295649385Z","level":"INFO","msg":"get RDS metrics"}                                                                                                                                                           │
│ {"time":"2024-03-26T12:04:17.294120478Z","level":"INFO","msg":"get RDS metrics"}                                                                                                                                                           │
│ {"time":"2024-03-26T12:04:47.29391695Z","level":"INFO","msg":"get RDS metrics"}                                                                                                                                                            │
│ {"time":"2024-03-26T12:05:17.293736005Z","level":"INFO","msg":"get RDS metrics"}                                                                                                                                                           │
│ {"time":"2024-03-26T12:05:47.294273412Z","level":"INFO","msg":"get RDS metrics"}                                                                                                                                                           │
│ {"time":"2024-03-26T12:06:17.294381841Z","level":"INFO","msg":"get RDS metrics"}                                                                                                                                                           │
│ {"time":"2024-03-26T12:06:47.294403924Z","level":"INFO","msg":"get RDS metrics"}                                                                                                                                                           │
│ {"time":"2024-03-26T12:07:17.294397064Z","level":"INFO","msg":"get RDS metrics"}                                                                                                                                                           │
│ {"time":"2024-03-26T12:07:47.294003217Z","level":"INFO","msg":"get RDS metrics"}  

[vmercierfr]

Great news!

We will adopt this approach for other parts of the exporter to make it more resistant.