Compare / Elastic
Obzly vs Elastic
Elastic Observability runs on Elasticsearch, a search engine, with Kibana as its UI. You can run it yourself on the free Basic tier, where SSO, audit logging and most alert connectors need a paid subscription, or use Elastic Cloud. Obzly stores logs, metrics and traces as Parquet in object storage and has no paid tiers.
Checked 27 October 2026 against Elastic's own docs and pricing. Sources
Which to choose
Choose Obzly if
- you want every signal in your own object storage without an Enterprise license for searchable snapshots;
- you want SSO, SCIM, an audit log and alerts to Slack or PagerDuty without a paid subscription;
- you'd rather run one image than an Elasticsearch cluster with data tiers, Kibana, Fleet Server and agents;
- you want a built-in MCP server without an Enterprise subscription;
- Apache 2.0 suits your legal team better than SSPL, AGPLv3 or ELv2.
Choose Elastic if
- you depend on full-text search across logs: Elasticsearch is a mature search engine, and ES|QL handles ad-hoc analysis;
- you want built-in machine learning: anomaly detection, log categorisation and forecasting (paid tiers);
- you need RUM, browser synthetic checks or eBPF continuous profiling (Enterprise), which Obzly doesn't have;
- you want a managed service: Elastic Cloud Serverless runs in 34 regions across AWS, Azure and GCP;
- you need SAML or Kerberos, which Obzly doesn't support, or Elastic's large catalogue of integrations and alert connectors.
What you run
Feature by feature
| Capability | Obzly | Self-managed (free Basic) | Paid tiers / Elastic Cloud | Source |
|---|---|---|---|---|
| License | Apache 2.0, every feature | SSPL, AGPLv3 or ELv2; default builds under ELv2 | Paid features need a subscription | Licensing FAQ |
| What you run | One image in up to five roles, Postgres, Valkey | Elasticsearch nodes, Kibana, Fleet Server, Elastic Agent or EDOT | Same, with ECK or ECE; nothing to run on Elastic Cloud | DeployFleet Server |
| Storage | ZSTD Parquet on S3, GCS, Azure Blob or MinIO | Indices on node disks; snapshots to S3, GCS or Azure | Searchable snapshots on object storage: Enterprise | Searchable snapshotsServerless differences |
| OTLP in | Logs, metrics and traces, gRPC and HTTP | All three, through an EDOT Collector or Agent you run | Managed OTLP endpoint on Elastic Cloud | Managed OTLPOTLP limitations |
| Query languages | Search bar and SQL for logs, PromQL, Jaeger API for traces | ES|QL, KQL, Lucene, Query DSL, EQL, SQL, PromQL | Same | Query languagesPROMQL command |
| Prometheus | Remote-write ingest and a PromQL API | Remote write 1.0 and PromQL, both GA since 9.5 | Same | Remote writePROMQL command |
| Dashboards | Own dashboards; imports Grafana layouts, not Kibana | Kibana; Grafana through a migration CLI, still evolving | Same | Migration platform |
| Alerting | Webhook, Slack, PagerDuty, email | Rules included; only Server Log and Index connectors | Email, webhook, Slack, PagerDuty, Teams and more | SubscriptionsConnectors |
| Synthetic checks | HTTP, TLS, TCP, DNS, ICMP from your probes; no browser | HTTP, TCP, ICMP and browser from private locations | Adds Elastic's managed locations (Cloud only) | Private locationsSubscriptions |
| Infrastructure inventory | Clusters to containers, plus AWS and Azure resources | Hosts, pods, containers and AWS resources | Same | Infrastructure inventory |
| RUM and profiling | Neither | RUM (User Experience); no profiling | Universal Profiling: Enterprise, not on Serverless | User ExperienceUniversal Profiling |
| SSO | OIDC with any provider; no SAML | Not available in Basic | SAML, OIDC, Kerberos, JWT: Platinum or Enterprise | SubscriptionsCloud SAML |
| SCIM | Included | Not supported | Not supported | Cloud SAML |
| RBAC and audit log | Per-project roles, teams, audit log with CSV export | Role-based access; no audit log | Audit log Gold and up; none on Serverless yet | SubscriptionsServerless differences |
| MCP server | Built in at /mcp, read-only | Not in Basic; the standalone server is deprecated | Agent Builder MCP server: Enterprise or Serverless | Agent Builder MCPSubscriptionsOld MCP server |
| Many regions | Control plane in front of many Sites | Cross-cluster search over remote clusters | Adds replication and cross-cluster ES|QL; none on Serverless | Cross-cluster searchSubscriptionsPricing |
| Price | Free; you run the infrastructure | Free; you run the infrastructure | Hosted from $99/month; Serverless per GB; self-managed via sales | PricingCloud HostedServerless |
Moving from Elastic
- Send to both. Add an OTLP exporter for Obzly to your EDOT or OpenTelemetry Collector next to the Elastic one. Collector guide
- Keep Prometheus flowing. Add Obzly as a second remote-write target. PromQL and Grafana
- Rebuild the dashboards you use. Obzly doesn't import Kibana dashboards. Recreate the panels with PromQL for metrics and search or SQL for logs.
- Rewrite alert rules. KQL and ES|QL rules become Obzly log searches or PromQL rules. Connect Slack, PagerDuty, email or a webhook. Alerts
- Move uptime monitors. HTTP, TCP and ICMP monitors map to Obzly checks from your own probes, which need a control plane. Browser monitors have no equivalent. Synthetics
- Switch off the Elastic cluster once Obzly has your full retention.
Sources
Every Elastic fact on this page comes from its vendor's own pages, read on 27 October 2026. Plan names and prices change; check the linked page before you decide.
- elastic.co/subscriptions: features by subscription tier
- elastic.co/pricing/faq/licensing: SSPL, AGPLv3 and ELv2
- elastic.co/pricing/cloud-hosted: Elastic Cloud Hosted prices
- elastic.co/pricing/serverless-observability: Serverless prices per GB
- Searchable snapshots docs: needs an Enterprise license
- Managed OTLP docs: Elastic Cloud only; self-managed uses a gateway
- Remote write docs: remote write 1.0 only
- Private locations docs: synthetics from your own network
- Universal Profiling docs: not on Serverless
- Migration platform post: Grafana dashboards to Kibana
Spotted something out of date? Open an issue.