It’s really awesome how this solves the “it’s hard to predict how many GB I’ll send”,
“It’s hard to find out what’s sending too much”, “It’s hard to predict how much retention I’ll get”, problems in one fell swoop!
Best for testing and small projects
Top Features
per month
Best for teams with production applications
Top Features
starts at
per month
Best for company-wide + large-scale applications
Top Features
Free | Pro | Enterprise | |
---|---|---|---|
Number of Users | Unlimited | Unlimited | Unlimited |
Number of Services | Unlimited | Unlimited | Unlimited |
Data Storage Limit | Unlimited | Unlimited | Unlimited |
Data Retention | 60-day | 60-day | 60-day (extensions available) |
Event Volume Limit | 20 million events per month | 100 million, 450 million, or 1.5 billion events per month | Scales to any size of business |
Honeycomb for Frontend Observability | Available as add-on | ||
Honeycomb Telemetry Pipeline | Available as add-on (10c per GB) | ||
Pricing | $0 per month | Starts at $130 / month or $1,300 / year | Contact us |
Subscription Period | Monthly | Monthly or Annually | Annually |
Distributed Tracing | |||
Triggers | 2 | 100 | 300 |
BubbleUp | |||
Team Query History | |||
Activity Logs | |||
Query Result Permalinks | |||
Slack Integration For Queries + Traces | |||
OpenTelemetry Support | |||
Agents, SDKs, Other Integrations | |||
Honeycomb Metrics | |||
Single-Sign On | |||
Service-Level Objectives (SLOs) | 2 | 100 | |
Service Map | |||
Refinery Dynamic Sampling | Enterprise support | ||
AWS PrivateLink | |||
Technical Support | Community support (Pollinators Slack) | ||
Honeycomb Support Response Times | N/A | Next business day | Two business hours |
Customer Success Team | |||
Onboarding Accelerator Packages | Available |
In Honeycomb, an event is a single trace span, structured log, or metric label combination that you send as OTLP or JSON to our API. This data can capture anything in your system that’s worth tracking, including all context fields that your application is instrumented to generate.
Each event can be up to 100KB, with up to 2000 attributes, and these attributes can have any number of values.
These events get analyzed in seconds whenever your team runs a query in Honeycomb.
Examples of a single event:
A span with multiple attributes:
// trace span
{ "Timestamp" : "2018-08-20T22:29:59.978735688Z" , "availability_zone": "us-east-1" , "build_id": 3150 , "customer_id": 1810 , "duration_ms": 448, "endpoint": "/api/v2/tickets", "fraud_dur": 112, "hostname": "app24" , "app.id": "6ee0d492cc3cbdc0", "is_error": false, "name": "/api/v2/tickets" , "platform": "android", "service_name": "api" , "status_code": 200, "trace.trace_id": "4be0d492cc3cbdc06ee0d492cc377dc0", "trace.span_id": adf691383eeb37368, "trace.parent_id": 097ace62b46dd221, "user_id": 483437 }
A narrow event with just a timestamp, host, and CPU metric:
// narrow event
{ "timestamp": "2022-06-05T14:44", "hostname": "a764" , "system.cpu.time": 192877 }
A wide event that’s a structured log:
// wide event
{ "Timestamp": "2022-07-01T19:23:50Z" , "cache_status": "HIT: , "client_ip_hash": "5c022873f02aac526c40013c12cd87c05745d82cd5dcebc74a0f307670f0b9f2" , "content_type": "binary/octet-stream", "downloaded_gem_name": "log-output-apm_metrics", "downloaded_gem_version": "2.0.1", "geo_city": "ashburn" , "geo_country_code": "US", "host": "rubygems.org", "is_h2": false, "is_ipv6": false, "is_tls": true, "log_resp_body_size": 3.9880235619286597, "log_time_elapsed": 3.141593230275788, "protocol": "HTTP/1.1" , "req_body_size": 0, "req_header_size": 174, "request": "GET", "request_accept_charset": "UTF-16", "request_accept_content": "text/html, image/gif, image/jpeg, *; q=.2, */*; q=.2", "request_accept_language": "*", "request_referer": "google.com", "request_user_agent": "Java/1.11.0_332", "resp_body_size": 9728, "resp_header_size": 755, "server_datacenter": "IAD", "service_id": "rubygems.org", "status": 200, "time_elapsed": 1386, "url": "/gems/log-output-apm_metrics-2.0.1.gem", "user_agent_name": "Java" }
Traces are comprised of multiple spans. Each span in a trace is counted as one event by Honeycomb. Each SpanEvent and each Link (in OpenTelemetry) count as one event.
From the Honeycomb web UI, find the Usage page under Team Settings. This page shows your team’s event volume for the last 60 days. In the table at the bottom of the page, you’ll also find the number of events received per dataset.
Team event limits are based on calendar months, regardless of the selected plan or when the team signed up. For example, if you started using Honeycomb with a billing day on the 10th of each month, all of your capacity management logic is still pinned to calendar month. The counter resets on the 1st of each month, even if your billing date is on the 10th. We use 30.4 days as the denominator for calculating daily ingest targets for events (eg. 100M EPM / 30.4 days = daily target of 3.29M events).
We send you a notification to let you know you have exceeded your monthly event limit. The goal of the notification is to give you time to either adjust your instrumentation down or upgrade to a bigger plan. If you have a second consecutive overage month, we send a notification that we’ll begin throttling your events to bring you down to your monthly event limit. You have 10 days from that notification to take remediation action before throttling begins.
We’ve added Burst Protection as a new measure to help account for unexpected spikes in traffic. Burst Protection automatically activates when you exceed your daily event target by at least 2X. With Burst Protection, any events in excess of the daily event target will not be counted against your team’s EPM limit. As an example, let’s say your team has a daily event target of 10 million events. One day, you see a spike in traffic and you send 30 million events to Honeycomb. Burst Protection automatically activates and the excess 20 million events sent that day will not be counted against your EPM limit. Burst protection can be triggered up to three times in a calendar month.
You receive a notification when Burst Protection is triggered. The effect of burst protection will also be visible on the Usage tab of your Team Settings page. While excess events under Burst Protection are not counted against your team’s EPM limits, they are considered successful events and they still appear in your per-dataset breakdown.
Event size limits are generous. Events are limited to 100kb size with a maximum of 2000 columns. Events over this size are rejected by Honeycomb with an error code in the API response. See our Events API documentation for more information.For reference, all of the text displayed on this FAQ page (including the intro and table of contents) is approximately 9kb in size.
Our data retention logic is based on time of ingest, so it will be there for 60 days from the date you sent it to Honeycomb (regardless of what’s on the timestamp of the events).
If you have an Enterprise plan, extended data retention is available on a per-month or per-yearly basis as an add-on. Pricing available upon request.
Honeycomb for Frontend Observability is available for enterprise customers and comprises of a software development kit (SDK) for OpenTelemetry’s browser instrumentation and the Web Launchpad within Honeycomb.
Honeycomb for Frontend Observability removes the limitations of legacy RUM products through easy OpenTelemetry-based instrumentation and support for hundreds of custom attributes at no cost. It also collects useful performance data like Core Web Vitals with attribution automatically, so you can easily identify what’s hurting your metrics.
All frontend data remains unaggregated and is immediately accessible in powerful visualizations that can be drilled into to surface insights into customer and performance issues. For example, attribution data for Core Web Vitals is automatically captured, so you can easily identify what exactly is hurting your Largest Contentful Paint metric. End-to-end tracing data is made available alongside performance data, to simplify debugging workflows.
Enterprise plans are annual subscriptions. Events Per Year (EPY) is a simple way to think about what is included for the duration of your plan. However, monthly targets for Enterprise plans are still calculated and set as EPM.
Enterprise plans start with a base allowance of 10 Billion Events Per Year. You can purchase additional capacity as needed with your annual subscription plan.
As your workload scales significantly, we offer volume discounts on additional EPY capacity to appropriately size an Enterprise plan that’s right for you.
Our enterprise ready security features include SSO, AWS PrivateLink, a SOC 2 Type II report, encryption at rest and in transit, ability to sign a BAA, regular penetration testing by an independent security firm, a bug bounty program, and GDPR compliance. Read more here.
Honeycomb is SOC 2 Type 2 certified and compliant.
For Enterprise plans or Pro plans on annual subscriptions, Honeycomb signs customer-provided DPAs. For Pro plans on monthly subscriptions, customers sign Honeycomb’s DPA.
Honeycomb can sign BAAs for customers. Honeycomb encrypts all data in transit and at rest, managed by AWS. We also offer AWS PrivateLink for clients that want to keep telemetry data within AWS and simplify configuration of outbound firewalls. Finally, we can offer guidance for preprocessing and normalizing data so that PHI/PII never leaves your infrastructure.
The table below outlines the various plans available, their EPM capacity, price, and discounts for annual purchase.
Plan | Up to Events per Month (EPM) | Cost per Month | Cost per Year |
---|---|---|---|
Free | 20M | $0 | $0 |
Pro | 100M | $130 | $1,300 |
Pro | 450M | $585 | $5,850 |
Pro | 1,500M | $1,950 | $19,500 |
Enterprise | Variable | n/a | Book a demo |
No, additional sales tax will be added to the monthly/annual bill in compliance with state guidelines and Honeycomb will remit them on your behalf to the appropriate entity. Please contact sales@honeycomb.io for more information.
Events are not cell phone minutes. Sorry, you can’t roll them over.
“It’s hard to find out what’s sending too much”, “It’s hard to predict how much retention I’ll get”, problems in one fell swoop!
Matt Button
Engineering Team Lead, Geckoboard
Storage is cheap and network/compute is expensive. Counting events vs GB, you want users to stuff all the data into their events so they can get the most out of it.
Julian Simioni
Co-Founder, Geocode Earth
Spencer Wilson
Senior Software Engineer, Optimizely
“It’s hard to find out what’s sending too much”, “It’s hard to predict how much retention I’ll get”, problems in one fell swoop!
Matt Button
Engineering Team Lead, Geckoboard
Storage is cheap and network/compute is expensive. Counting events vs GB, you want users to stuff all the data into their events so they can get the most out of it.
Julian Simioni
Co-Founder, Geocode Earth
Spencer Wilson
Senior Software Engineer, Optimizely