Prompt
I hate how Datadog gets expensive when logs spike, what are the alternatives?
Latest observation
Oct 19, 2026
OpenAI APIWeb search: off
If your pain is Datadog logs getting expensive during spikes, the main thing to optimize for is usually pricing model + ingestion control. A lot of alternatives are cheaper, but they differ a lot in searchability, retention, and operational effort.
Good alternatives to consider
1) Grafana Loki
- Best for: cheaper log storage, especially if you already use Grafana/Prometheus
- Why it’s cheaper: indexes far less than ELK-style systems
- Tradeoff: less powerful full-text search than Datadog/Elastic
- Typical setup: Loki + Grafana + Promtail/Fluent Bit
2) Elastic / OpenSearch
- Best for: powerful search and analytics
- Why: very flexible, mature ecosystem
- Tradeoff: can get expensive/operationally heavy if you run it yourself; managed versions can still cost a lot
- Good if: you need rich querying, dashboards, and are okay tuning index/retention policies
3) Self-hosted OpenSearch
- Best for: lower-cost control over infrastructure
- Why: open source, no vendor log tax
- Tradeoff: you own scaling, tuning, upgrades, and cluster health
- Good if: you have infra expertise and want to cap spend
4) Cloud provider logging
Examples:
- AWS CloudWatch Logs
- GCP Cloud Logging
- Azure Monitor Logs
- Best for: simplicity if you’re already on that cloud
- Tradeoff: can also become expensive at scale, and querying UX varies
- Good if: you want “good enough” and fewer moving parts
5) Better Stack / Logtail
- Best for: simpler SaaS log management
- Why: often more predictable and cheaper than Datadog for logs
- Tradeoff: not as broad as Datadog for APM/infra/metrics all-in-one
- Good if: you mainly want logs with decent UX and sane pricing
6) Coralogix
- Best for: log analytics with more cost control options
- Why: often priced around ingestion/storage/usage differently than Datadog
- Tradeoff: still a commercial SaaS; evaluate total cost carefully
- Good if: you want enterprise features without Datadog’s billing style
7) Humio / CrowdStrike Falcon LogScale
- Best for: high-volume logs
- Why: designed for fast search at scale
- Tradeoff: may be overkill; pricing still needs scrutiny
- Good if: you have large bursts and need fast incident investigation
Ways to cut cost without fully replacing Datadog
If you don’t want a migration, you can still reduce the spike bill:
- Sample noisy logs
- Drop debug logs in prod
- Route only important logs to Datadog
- Use short retention
- Filter at the agent/collector
- Send raw logs to cheap storage like S3 and only index a subset
- Aggregate repetitive logs
- Use metrics for counts instead of logs for everything
A common pattern is:
- Hot searchable logs in Datadog/Loki/Elastic
- Cold archive in S3/GCS
- Alerts from metrics, not log volume
What I’d recommend by scenario
- Need cheapest + okay with less search power: Grafana Loki
- Need strong search and don’t mind ops: OpenSearch/Elastic
- Want SaaS with more predictable costs: Better Stack or Coralogix
- Need massive log volume search: LogScale
- Just want fewer surprises and use AWS/GCP/Azure already: cloud-native logs
Quick rule of thumb
- If your main issue is “Datadog bills me for volume spikes”, look first at Loki or Better Stack
- If your main issue is “I need powerful log search”, look at Elastic/OpenSearch
- If your main issue is “I want to stop paying for every byte”, implement log filtering/sampling + cold storage
If you want, I can give you a ranked shortlist based on your stack — for example: AWS vs GCP, Kubernetes vs VMs, and whether you need APM/metrics too.