Independently operated. Not affiliated with Datadog, New Relic, Grafana Labs, Dynatrace, Splunk, or Elastic. Pricing sourced from public pages and may not reflect current rates. Verify on each vendor's pricing page before purchasing.
MonitoringCost.com
Independent monitoring-cost research, read by teams pricing their stack.Sponsor this site →

Azure

Azure Application Insights pricing 2026

Verified September 2026

Application Insights has no price of its own. Its telemetry bills on the Log Analytics workspace underneath it, with one exception that saves most teams their entire retention bill: Application Insights tables keep 90 days of data free, not 31. Here is every meter, what the availability tests cost, and where sampling actually bites.

How much does Application Insights cost?

There is no Application Insights per-GB rate. A workspace-based resource writes telemetry into a Log Analytics workspace and bills at $2.30 per GB of Analytics Logs ingestion (after 5 GB free per billing account per month). Retention is where Application Insights differs: its tables get 90 days free, not the 31 days that applies to Analytics Logs generally, so most teams pay $0 for retention. Past day 90 it is $0.10/GB/mo. Standard availability tests bill separately at $0.0005 per execution, and Live Metrics Stream is free. Commitment tiers apply, taking ingestion to $1.96/GB at 100 GB/day. East US list rates checked 26 September 2026.

TL;DR

Pricing Application Insights means pricing Log Analytics: $2.30/GB ingestion, 5 GB free per billing account per month. The one Application Insights-specific break is retention, where its tables keep 90 days at no charge against 31 days elsewhere. Availability tests are a real and easily-missed line: $0.0005 per execution means one URL at five-minute frequency from five locations costs $21.60/month before any telemetry charge. URL ping tests retire 30 September 2026. Sampling, not plan choice, is the lever: there is no cheaper Application Insights plan to move to.

The model

Application Insights does not have a price

The question "what does Application Insights cost per GB" has no direct answer, and that is the first thing to understand about the bill. A workspace-based Application Insights resource is a view over data sitting in a Log Analytics workspace. Microsoft is explicit that billing for ingestion and retention is done by the workspace where the data lives, and that all the Log Analytics pricing options, including commitment tiers, are available as a result. There is no separate Application Insights data meter to look up.

In practice that means application telemetry lands on the same Analytics Logs meter as infrastructure logs: $2.30 per gigabyte in East US, after a free allowance of 5 GB per billing account per month. That allowance is shared across the whole billing account rather than granted per workspace or per app, so a second application does not get its own 5 GB.

It also means application telemetry counts toward a commitment tier. A workspace on the entry 100 GB per day tier ingests its Application Insights data at $1.96 per gigabyte rather than $2.30, and the app telemetry helps reach the tier alongside everything else in the workspace. Two workloads that would each miss the threshold alone can clear it together. The full tier ladder is on our Azure Monitor cost breakdown.

Every meter

What each Application Insights line item costs

East US pay-as-you-go list pricing, read on 26 September 2026 from Microsoft's Azure Retail Prices API and Microsoft Learn (Azure Monitor Logs cost). Azure Monitor prices are set regionally; other regions differ.
MeterRateFree / includedHow it bills
Telemetry ingestion (via the workspace)$2.30/GBFirst 5 GB/billing account/monthRequests, dependencies, traces, exceptions, page views. Billed on the Log Analytics Analytics Logs meter, not an Application Insights meter
Analytics retention$0.10/GB/monthFirst 90 days free on Application Insights tablesOnly data held past day 90 bills. Configurable up to two years
Long-term retention$0.02/GB/monthNoneArchive up to 12 years; a search job bills per GB scanned to retrieve
Standard availability test$0.0005/executionNonePer execution, per location. Test telemetry is also billed as ingestion
Multi-step web test$10/monthNoneDeprecated by Microsoft. Standard tests are the supported path
Live Metrics StreamNo chargeAlwaysMicrosoft states there is no data-volume charge for Live Metrics Stream
Legacy Per Node (Enterprise)$15/node/month200 MB/node/day includedOperations Management Suite E1/E2 customers only; overage bills at $2.30/GB

The 90-day break

Application Insights retention is free for 90 days, not 31

This is the most widely misstated fact in Application Insights pricing, and it is worth several thousand dollars a year to a mid-size estate. The 31-day figure that circulates is correct for Analytics Logs in general: 31 days of analytics retention is included in the ingestion price, and Microsoft notes that dropping a table below 31 days does not reduce cost because you have already paid for it.

Application Insights tables are not in that group. Microsoft's retention documentation carries a section headed "Log tables with 90-day default retention" which states that tables related to Application Insights resources keep data for 90 days at no charge, and lists them: AppRequests, AppDependencies, AppTraces, AppExceptions, AppEvents, AppMetrics, AppPageViews, AppPerformanceCounters, AppBrowserTimings, AppAvailabilityResults and AppSystemEvents. The Azure Monitor pricing page states the same allowance from the billing side, for both classic and workspace-based resources.

The practical consequence is that a team retaining a quarter of application telemetry has a retention bill of exactly zero, and only data held past day 90 touches the $0.10 per gigabyte per month meter. Each of those tables can also be adjusted individually, so you can hold AppRequests for a year while letting AppTraces expire at 90 days, without paying retention on the traces.

Where this comes from

Microsoft Learn, Manage data retention in a Log Analytics workspace, section "Log tables with 90-day default retention", and the Azure Monitor pricing page, which lists 90 days free for data ingested into either classic or workspace-based Application Insights. Both read 26 September 2026.

Availability tests

The web test meter people forget to budget

A standard availability test bills at $0.0005 per test execution in East US, and the result telemetry is billed as ingestion on top. The charge scales with frequency multiplied by locations, which is why it surprises people: adding locations to make alerting reliable multiplies the bill by the same factor.

Work it through on Microsoft's own recommended configuration. A five-minute test frequency from five locations is 12 executions per hour per location, so 12 x 24 x 30 x 5 = 43,200 executions a month, which at $0.0005 is $21.60 per URL per month. Ten monitored endpoints on that schedule is $216.00 a month before a byte of telemetry is counted. Microsoft recommends a minimum of five locations, allows up to 16, and caps you at 100 availability tests per Application Insights resource.

Two deprecations matter here. Multi-step web tests still carry a $10 per month meter but Microsoft lists them as deprecated. More urgently, URL ping tests retire on 30 September 2026 and existing tests are removed from resources on that date. Anyone still running them has to move to standard tests, and Microsoft's own migration guidance is blunt that this is a budget event: it warns that standard tests incur charges when enabled and tells you to review Azure Monitor pricing before turning the migrated tests on. A large estate of ping tests should be costed at $0.0005 per execution before it converts, not after.

How the $21.60 is built

Our arithmetic over one published rate. 60 minutes divided by a 5-minute frequency is 12 executions per hour per location; 12 x 24 hours x 30 days x 5 locations = 43,200 executions; 43,200 x $0.0005 = $21.60. The $0.0005 Standard Web Test Execution rate is East US list pricing from the Azure Retail Prices API. Test frequency, location guidance and the 30 September 2026 URL ping retirement are from Microsoft Learn (availability tests).

Worked examples

What Application Insights costs at three volumes

These are our own arithmetic over the published East US rates above, not quoted prices. Every input is on the page, so you can redo them with your own volumes.

A small app sending 3 GB a month. Ingestion falls entirely inside the 5 GB per billing account free allowance, so it is $0. Retention is inside the 90-day free window, so that is $0 too. The total is nothing, provided nothing else in the billing account has already consumed the 5 GB. That caveat is the whole catch: the allowance is per billing account, so the second app in the same account usually finds it already spent.

A mid-size app sending 60 GB a month, 90-day retention. Ingestion is (60 - 5) x $2.30 = $126.50 a month. Retention is free because 90 days is the included window. Total $126.50 a month.

The same app retaining 12 months. Ingestion is unchanged at $126.50. At steady state you hold 720 GB, of which the most recent 90 days (180 GB) is free, leaving 540 GB on the retention meter at $0.10 per gigabyte per month, or $54.00. Total $180.50 a month. Note what the 90-day allowance is doing: on the commonly quoted 31-day assumption you would expect to pay retention on roughly 660 GB, and budget about $12 a month more than you owe.

A high-volume app sending 400 GB a month. That is about 13 GB a day, well under the 100 GB per day commitment floor, so it stays on pay-as-you-go: (400 - 5) x $2.30 = $908.50 a month. Reaching the entry commitment tier takes roughly 3,000 GB a month across the whole workspace, which is why app telemetry is usually pooled with infrastructure logs rather than given its own workspace.

Where it bites

Three things that inflate an Application Insights bill

Sampling switched off

Adaptive sampling is on by default in the ASP.NET and ASP.NET Core SDKs and in Azure Functions, targeting five telemetry items per second per host. Teams that disable it while debugging and never re-enable it, or that run Java (where sampling is off by default in the SDK), pay the full unsampled ingestion rate indefinitely.

Availability tests times locations

$0.0005 an execution looks like nothing until you multiply. Fifty endpoints at five-minute frequency from five locations is $1,080 a month in execution charges alone, and the URL ping retirement on 30 September 2026 pushes estates onto this charged meter whether or not they budgeted for it.

Chatty dependency telemetry

Dependency tracking records every outbound call, so a service that talks to a cache or a database in a tight loop can generate more AppDependencies volume than everything else combined. It is the first thing to put behind a sampling override, and in Java that is exactly what sampling overrides are for.

Cost levers

Three things to do about Application Insights cost

Sample, because there is no cheaper plan

Ingestion is the bill and sampling is close to a linear lever on it. Halving ingested volume takes the 60 GB example from $126.50 to (30 - 5) x $2.30 = $57.50 a month. Adaptive sampling is the default in the .NET SDKs; fixed-rate lets you pick the number and keeps client and server in step. Metrics and performance counters are never sampled, so dashboards and metric alerts stay accurate.

Pool into a commitment tier

Application Insights data counts toward the workspace commitment tier like any other Analytics Logs ingestion, so app telemetry and infrastructure logs in one workspace can reach a tier neither would reach alone. The entry 100 GB per day tier takes the effective rate from $2.30 to $1.96 per gigabyte.

Tune retention per table, not per workspace

The 90-day allowance is free, so there is nothing to save inside it. Past that, retention is set per table: hold AppRequests for a year if the business needs the trend, and let AppTraces and AppDependencies expire at 90 days rather than paying $0.10 per gigabyte per month to keep debug noise nobody queries.

Verify before you commit

Every rate on this page is East US pay-as-you-go list pricing read from the Azure Retail Prices API on 26 September 2026. Billing mechanics are from Azure Monitor Logs cost calculations and options, retention from Manage data retention, sampling behaviour from Telemetry sampling in Application Insights, and the legacy Per Node mechanics from the legacy enterprise (per node) pricing tier. Azure Monitor prices are set per region and Enterprise Agreement discounts apply on top; confirm your own region before committing.

Frequently asked

How much does Azure Application Insights cost?
Application Insights has no per-gigabyte rate of its own. A workspace-based resource writes its telemetry into a Log Analytics workspace and bills on that workspace's meters: $2.30 per gigabyte of Analytics Logs ingestion in East US, after a free allowance of 5 GB per billing account per month. The difference that matters is retention. Application Insights tables keep data for 90 days at no charge, where general Analytics Logs tables get 31, so most teams pay nothing at all for retention. Past day 90 the rate is $0.10 per gigabyte per month. On top of ingestion, standard availability tests bill at $0.0005 per test execution and Live Metrics Stream is free. Rates are East US list prices from the Azure Retail Prices API, checked 26 September 2026; Azure Monitor prices are set regionally.
Does Application Insights really get 90 days of free retention?
Yes, and it is the single most misreported fact about Application Insights pricing. Microsoft's retention documentation lists the Application Insights tables (AppRequests, AppDependencies, AppTraces, AppExceptions, AppEvents, AppMetrics, AppPageViews, AppPerformanceCounters, AppBrowserTimings, AppAvailabilityResults and AppSystemEvents) under a heading titled "Log tables with 90-day default retention" and states that they keep data for 90 days at no charge. The Azure Monitor pricing page says the same thing: 90 days free for data ingested into either classic or workspace-based Application Insights. The 31-day figure that gets quoted is the allowance included in the Analytics Logs ingestion price generally, and it does not apply to these tables. A team retaining a quarter of application telemetry has a retention bill of zero.
What do Application Insights availability tests cost?
A standard availability test bills at $0.0005 per test execution in East US, and the test's own result telemetry is billed as ingestion like any other data. Cost therefore scales with frequency times locations, not with the number of URLs alone. One URL tested every five minutes from five locations runs 43,200 executions a month, which is $21.60 in execution charges; ten URLs on the same schedule is $216.00. Microsoft recommends a minimum of five test locations and allows up to 16, and you can create up to 100 availability tests per Application Insights resource. Multi-step web tests bill at $10 per month but are deprecated. URL ping tests retire on 30 September 2026 and existing ones are removed from your resources, so single-step checks have to move to standard tests. Microsoft's migration guidance warns in terms that standard tests incur charges when enabled and tells you to review Azure Monitor pricing before switching them on, so treat the conversion as a budget item rather than a like-for-like swap.
Is there a cheaper Application Insights pricing tier?
No. There is no cheaper plan to move to, because there is no plan: workspace-based Application Insights bills on whatever the underlying Log Analytics workspace bills. That leaves three real levers. Ingest less, through adaptive or fixed-rate sampling in the SDK. Reach a commitment tier, which applies to Application Insights data like any other Analytics Logs ingestion and takes the effective rate from $2.30 to $1.96 per gigabyte at 100 GB per day. And route verbose custom telemetry into a Basic Logs table at $0.50 per gigabyte where the schema allows it. Changing tier is not on the list because there is no tier to change.
How does Application Insights sampling reduce the bill?
Sampling reduces the volume of telemetry that reaches the workspace, and ingestion is what you pay for, so it is close to a linear lever. Adaptive sampling is on by default in the ASP.NET and ASP.NET Core SDKs and in Azure Functions, targeting five telemetry items per second per server host; it does not drop anything until that rate is reached, so low-traffic apps are typically not sampled at all. Fixed-rate sampling lets you set the percentage yourself and synchronises client and server so related page views and requests stay together. Ingestion sampling runs at the Azure endpoint and reduces what is processed, retained and charged for, but not the network traffic leaving your app. Metrics, custom metrics and performance counters are never sampled by any of the three methods.
What is the legacy Application Insights Per Node pricing tier?
It is the old Enterprise tier, renamed Per Node, and it is visible only to customers with an Operations Management Suite E1 or E2 subscription. It lists at $15 per node per month in East US, charged only for hours in which a node actually sends telemetry, at the monthly rate divided by 744. Each node contributes a 200 MB per day data allowance to a subscription-wide pool, and ingestion beyond that pool bills at $2.30 per gigabyte. Microsoft's own worked example: four nodes sending telemetry for 15 of 24 hours give ((4 x 15) / 24) x 200 MB = 500 MB included for that day, so 1 GB of telemetry that day costs $1.15. Microsoft recommends against it, warning that tracking usage in this tier is complex because all resources in a subscription report against just one of them.
Do classic Application Insights resources cost the same?
Classic resources are retired, and Microsoft's documentation now refers to them in the past tense. For anything still running on one, the billing answer is that data ingestion and retention follow the same pay-as-you-go rates as workspace-based resources, but classic resources cannot use commitment tiers at all. That is the clearest cost argument for converting: a subscription with enough volume to reach the 100 GB per day tier is paying $2.30 per gigabyte on a classic resource where a workspace-based one would pay $1.96.

Didn't find your answer?

Ask us. A real person reads every question and we answer the ones we can, with sources. If your question would help other readers, we may publish an anonymised version, with your permission. General reference only.