Learning the tricks of Loki: Distributed logging at scale
Anais Urlichs
SRE @CivoCloud
Alex Jones
Principal Engineer @CivoCloud
What is Civo?
Under 90s launch times
Transparent Pricing
Cloud-Native first
Introduction to Loki
Demo
The tricks of Loki
Agenda
Why do we use Loki?
Traditional logging solutions are no longer fit for purpose.
Expensive
Difficult to operate at scale
Challenging to correlate against metrics
Loki
We want a cost-effective way to collect, store and query logs.
Loki is a highly available, horizontally scalable, multi-tenant log aggregation system.
Cost-effective
Scales with your infrastructure
Integrates with your observability stack
Promtail
Logs
Logs
Promtail
Discover local logs
Process logs
Ship processed logs to Loki
How do we use Loki?
syslog
Compute�Node
Kubernetes�Control plane
Regional Supercluster
Pods
Promtail
Distributed�Gateway
Ingestor
Object�Storage
Query/Frontend
Querier
Index Memcache
Demo
The Tricks of Loki
Loki can be configured to suit your infrastructure topology and ingestion needs.
Results Memcache
Index Memcache
Chunks Memcache
Caching
Scaling
Distributor
Query/Frontend
Querier
KV stores
S3
GCS
Cassandra
And more....
Combine Loki with Tempo to extract traces from your logs.
Labels are powerful shortcuts to your logs.
LogQL takes querying to the next level.
rate({app="civo-csi-controller",namespace="kube-system",pod="civo-csi-controller-0"} |~ "error" [5m])
Filter Expression
Log Selector
Loki enables the creation of powerful visualisations
What’s next for Civo?
Inter-regional logging
Supply Chain observability
Additional Synthetic Monitoring for tenant clusters
Q & A
Thank You!
@urlichsanais
@AlexJonesax