bomonike

Prometheus logs, metrics, traces faster into a more compact database with built-in GUI instead of Grafana

openobserve.ai (“O2”) is an “Open source observability platform for logs, metrics, traces, frontend monitoring, pipelines and LLM observability.”

g2.com/products/openobserve/reviews shows high 5/5 ratings. Put in a word if you’re familiar with it.

Here’s what makes OpenObserve better and cheaper than Datadog, Splunk, Elasticsearch, Prometheus:

Isaac Johnson presented his large deployment costs vs. competitors GroundCover, DataDog, New Relic:
2026-07-openobserve-39.png
OpenObserve cost $255/mo for 250 GB of logs, 250 GB metrics, 10 GB traces with 30 hosts and 100 users,

Pricing: Enterprise Edition is free continually for up to 50 GB/day (~1.5 TB/month) of data ingestion. But you can exceed your license limit up to 3 times per month before being blocked.

Query is $0.01 per GB with 30-Day Non-metric (Logs, Traces) retention and 15-Month Metrics Retention.

Architecture

  1. What does OpenObseve (O2) provide over a traditional observability stack based on Prometheus/Grafana?

    o2-arch-before-1890x1356.png

  2. Instead of a separate Query Language design can run into bottlenecks:
    • Loki — bottlenecks on ingestion streams and chunk storage accessed using LogQL language
    • Tempo — bottlenecks on trace storage and block compaction accessed using TracQL language
    • Mimir — bottlenecks on series cardinality and query concurrency

  3. These potential bottlenecks are addressed by the new architecture of OpenObserve’s Unified Agent Ingestion. o2-arch-2258x1372.png

  4. Instead of a separate Grafana process to manage, OpenObserve provides its own built-in UI, all in a Kubernetes cluster created by a single Helm chart.
  5. Within the cluster, OpenObserve is able to lower storage costs because it uses a Parquet-format Apache Arrow columnar database

The company

VIDEO: Founder Prabhat Sharma revealed that OpenObserve was rewritten from Go into Rust for higher compression (~40x) using columnar storage with Apache Arrow Parquet SQL to achieve 140x lower storage cost vs. Elasticsearch. * https://www.linkedin.com/company/openobserve/ * https://openobserve.ai/ * https://openobserve.ai/docs/

At time of this writing, their “The Observer” newsletter had 1,737 subscribers from 6,181 followers DelRel:

OpenObserve is open-sourced with a AGPL 3.0 license for local install from:

Quickstart demos

There are three ways to obtain a demo environment installer with credentials baked in, and install locally on a Mac:

  1. Install Helm on your local machine being observed. 
    brew install helm
helm version

    version.BuildInfo{Version:"v4.2.3", GitCommit:"43e8b7feece8beb0fcba47059ec9b522fd929a64", GitTreeState:"clean", GoVersion:"go1.26.5", KubeClientVersion:"v1.36"}
  2. Install OpenObserve demo locally: A. Docker demo image download
    B. curl and run
    C. cloud run

The software is localized in 11 languages: English, German, French, Italian, Japanese, Korean, Dutch, Portuguese, Spanish, Turkish, Chinese (Simplified)

If you have the demo environment already running, go to Run Demo.

A. Docker demo image download & install

  1. Install Docker Desktop, if you don’t already have it.
  2. Start Docker Desktop and configure for Kubernetes.
  3. Download the Docker image: 
    docker pull public.ecr.aws/zinclabs/openobserve:v0.91.2
  4. Optionally, get the disk space used by the image: 
    docker image ls | grep openobserve

    Observe the “309MB” in the response (at time of writing):

    public.ecr.aws/zinclabs/openobserve:v0.91.2        a39535f64535        309MB             0B   U
  5. To run using default email, password, and configurations for a demo baked into the Docker image: 
     docker run -v $PWD/data:/data -e ZO_DATA_DIR="/data" -p 5080:5080 -e ZO_ROOT_USER_EMAIL="root@example.com" -e ZO_ROOT_USER_PASSWORD="Complexpass#123" public.ecr.aws/zinclabs/openobserve:v0.91.2
 20
  6. Go to Run Local Demo.

B. curl and run demo locally

  1. Create and navigate to a folder to receive the download, such as “o2”.

  2. Switch to an internet browser at: https://openobserve.ai/downloads

  3. Select Edition: “Enterprise” and Deployment Mode: “Single Node”.
  4. Click operating system “MacOS”.
  5. If you want to confirm valid download, double-click the SHA256 UUID to copy to your Clipboard and switch to a CLI Terminal to save it in a file: 
    export SHA1=$(pbpaste)

  6. Click the copy icon for the curl command to capture the Quick Install Script into your Clipboard.

    REMEMBER: There is not “brew install openobserve” available instead.

    curl -L https://raw.githubusercontent.com/openobserve/openobserve/main/downloadO2.sh | sh -s o2-enterprise v0.91.3

    Observe that the latest version available (v0.91.3 at the time of this writing) has been automatically updated from https://github.com/openobserve/openobserve/releases/

       % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                   Dload  Upload   Total   Spent    Left  Speed
 100  1445  100  1445    0     0   5023      0 --:--:-- --:--:-- --:--:--  5034
 Detecting platform...
 Platform: darwin
 Detecting architecture...
 Architecture: arm64
 Downloading: https://downloads.openobserve.ai/releases/o2-enterprise/v0.91.3/openobserve-ee-v0.91.3-darwin-arm64.tar.gz
   % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                   Dload  Upload   Total   Spent    Left  Speed
 100  124M  100  124M    0     0  10.8M      0  0:00:11  0:00:11 --:--:-- 11.4M
 Extracting...
 ✅ Download and extraction complete!
  7. Optionally, generate a SHA256 for the downloaded file and compare versus the previous SHA saved to variable SHA1 : 
    export SHA2=$(shasum -a 256 openobserve | awk '{print $1}')
diff <(echo "$SHA1") <(echo "$SHA2")
  8. Run openobserve for demo purposes using default credentials within the file: 
    export ZO_ROOT_USER_EMAIL=root@example.com
export ZO_ROOT_USER_PASSWORD=Complexpass#123
./openobserve

    SECURITY PROTIP: The above exposes secrets. Instead, during regular/production usage, to protect your password, create a shell file to run OpenObserver after looking up passwords and other secrets from a secrets vault and peform the export.

  9. Click “Allow” to the pop-up message “Do you want the application “openobserve” to accept incoming network connections? Clicking Deny may limit the application’s behavior. This setting can be changed in the Firewall pane of Network Settings.”

  10. Go to Run Local Demo.

C. cloud run your own account

OpenObserve holds its users in several geographic ares within two cloud providers (AWS and Azure). So open a cloud account in AWS and/or Azure using the email you will use with OpenObserve.

  1. Switch to an internet browser at: https://openobserve.ai/downloads

  2. Click “Log In” to select a geographic area within a cloud vendor (AWS or Azure). o2-geo-clouds.png

  3. Provide your cloud email, password, 2FA, and consent for the cloud you selected.
  4. If you selected Azure, you’ll arrive at https://us2.openobserve.ai/web

  5. Click “Start your 14-day Trial”. Note the date in your secrets vault utility.

    Observe that no credit card is requested before a subscribing.

  6. Go to Run Local Demo.

Run Local Demo

  1. Switch to an intenet browser to URL: http://localhost:5080/web/login

  2. Double-click the default User Email and Password baked into the installer and paste to Login:
    • User Email: root@example.com
    • Password: Complexpass#123
  3. Follow instructions illustrated at
    here
    https://freshbrewed.science/2026/07/16/openobs.html#windows-logs

Configure Dashboard

When you’re ready to use your own account:

  1. Look at the documentation website:
    https://openobserve.ai/docs

At the OpenObserve dashboard:

  1. Click the user icon at the upper-right to “Manage Theme”.

  2. Select “Dark”, then “O2 Signature” or your Custom Color. Click “X” to exit.

  3. Menu items “AI”, “Incidents”, and “Billing” may not appear in your menu.

    Organization, Users

  4. Click the “default” pull down at the top-right. Observe that the URL of the page is the “org_identitifier=” string.

  5. Bookmark the URL with the org_identifier.

  6. Click “IAM” at the bottom of the left menu.
  7. Click “Organizations” to optionally click the pencil icon to change from “Default” to your preferred name.

  8. Optionally, click “New organization”.

  9. Add User Groups, Users, Roles, Invitations for each organization.

  10. CLick “Create New Token” to create an “Ingestion Tokens” for each injestion service.

  11. Click “Settings” at the bottom of the left menu.

  12. Here is where you can delete an organization.

  13. Click on the left menu any item (“Logs, Metrics, Traces, AI, RUM, Dashboards, Alerts, Incidents, Data”) for “Data sources” to appear.
  14. Configured Kubernetes to run within three major operating sytems:
    • Windows
    • Linux
    • macOS
  15. O2 also runs within the three major clouds:
    • Amazon Web Services (AWS)
    • Google Cloud Platform (GCP)
    • Microsoft Azure
  16. Configure Traces with OpenTelemetry spans

  17. Configure Real User Monitoring (RUM)

    Data Sources

    o2-data-menu-1051x91.png

    These were identified in the Architecture diagram above and documentation at
    https://openobserve.ai/docs/ingestion

  18. Configure Data Source for macOS. Click the icon to get the code to “Install the Agent”: 
    curl -sSL https://raw.githubusercontent.com/openobserve/o2-datasource/main/k8s/install.sh | bash -s --    --cluster-name=cluster1   --o2-url=http://localhost:5080   --org-id=default   --access-key=...

    ✓ cert-manager installation initiated
ℹ Waiting for cert-manager webhook to be ready (timeout: 300s)...
  19. Click “Test” button to “Verify Data in OpenObserve”.
  20. Click “Dashboards” menu, “New Dashboard”, named “???.
  21. Click “Add panel”. TODO:

Unified Log setting

The unified log is high volume. To reduce it, set LEVEL or PREDICATE in

   /opt/openobserve-collector/macos-unified-log.sh

and run:

   sudo launchctl kickstart -k system/ai.openobserve.macos-unified-log

Specific O2 Features

  1. Click “OpenObserve Features” for this pop-up:

    OpenObserver provides access to 30+ prebuilt dashboards to kickoff your observability strategy.

VRL

VRL (Vector Remap Language) is a stateless scripting language initiated by Datadog’s open-source team for transforming and processing observability data (logs, metrics, and traces): parse, filter, enrich, or reshape telemetry events before they reach storage or dashboards.

https://github.com/vectordotdev/vrl is built using Rust as a https://crates.io/crates/vrl

Play on the vrl.dev Playground

https://github.com/vectordotdev/vrl/blob/main/examples/simple.rs

Tutorials

TODO:

References

TODO:

v009 + G2 @openobserve.md created 2026-07-26