Multiple Web apps on a App service plan

Question

Multiple Web apps on a App service plan

ReCloudS 80

Hi ,

I have a question on monitoring regarding on the Azure App service plan , my question is I have deployed 3 web apps and 3 function apps on azure app service plan which the service plan is P2V3 and the instance count is 3 meaning the each app/fun is getting shared on all the 3 instances when am checking the metrics for the app service plan , the max it says is ~60-70% for the time period of 30 days , when am checking for sum it was showing as ~1000-1200% for 30 days , it means that the whole apps are consuming what they got allocated. And I suspect that one web app is consuming more memory as I also observed the critical error for the memory in diagnose and solve problems of that web app , so I need a suggestion here whether am on the right way of looking at it ? if yes ? whats the suggestion to isolate the app with diff app service plan or shall I just increase the instance count to even more ? and also help me with how much the apps are consuming how much memory? User's image

User's image

Accepted answer

1 additional answer

Your answer

Answer 1

Bhargavi Naragani 1,690 Microsoft External Staff

Hi @ReCloudS
In addition to @Marcin Policht response, and after addressing your concern.

You are correct, under a P2V3 configuration (16GB RAM per instance, 3 instances equaling 48GB), memory shortages shouldn't normally occur. However, even if one app might consume up to 9GB while the rest consume considerably less, there is still room for instance-level bottlenecks if memory usage isn’t evenly distributed.

Isolate the High-Usage App, transfer it to a different App Service Plan to prevent affecting other apps. Create additional instances (e.g., scale from 3 to 4 or 5) to better distribute memory load. Use Application Insights to look into memory usage and fix leaks or inefficiencies. Use Azure Monitor's Memory Working Set metric or Kudu's Process Explorer to monitor per-app memory usage.

Regarding Cost Explanation Yes, cost is an instance basis, if one instance costs $1000/month, three instances equal $3000, plus data transfer charges, as applicable.

https://azure.microsoft.com/en-us/pricing/details/app-service/linux/
https://learn.microsoft.com/en-us/azure/app-service/monitor-app-service
https://learn.microsoft.com/en-us/azure/app-service/troubleshoot-diagnostic-logs

If the answer is helpful, please click Accept Answer and kindly upvote it so that other people who faces similar issue may get benefitted from it.

Let me know if you have any further Queries.

ReCloudS 80 Reputation points

2025-03-07T06:25:17.2266667+00:00

@Bhargavi Naragani Thank you got an better idea now how those web apps work under app service plans , also can you help me how do I take an average for the monitoring metrics regarding the app service plan and seperately for the web apps or function apps ? like for last 30 days or 24 hrs how do I consider if I need to check for the usage of the web apps/function apps under the app service plan meaning do I need to also check for the whole app service plan ? for example If my app service plan is p0v3 with 2 instance count and holds 9 web apps which consume hardly woeking memory set max is 200MB and average 40MB for 30 days which i checked for , and for the app service plan I when I check for the memory and CPU for 30 days which the max was getting 100% CPU and 87% memory consuming on the App service plan level ,is this way that I can make consideratons for consumptions and make changes ? or do I have to take an average or max for 24hrs or last 7 days , do I need to consider both Average and Max for that ?
Bhargavi Naragani 1,690 Reputation points Microsoft External Staff

2025-03-07T11:36:49.78+00:00

Hi @ReCloudS,

Monitoring Metrics for App Service Plan vs. Web Apps/Function Apps, App Service Plan Metrics reflect the total resource usage across all apps sharing that plan.

If CPU Percentage often reaches 100%, your instances are heavily loaded and might require scaling. High memory usage (such as 87%) indicates potential resource contention, an app or apps might be using too much memory.

Per App Metrics assist in determining specific apps consuming more resources. Check both Average and Max values within your selected time interval (such as 24 hours, 7 days, or 30 days) Using Average reflects typical behavior over time and informs you of steady load and using Max Indicates peak usage and possible stress points. CPU Time Assist in monitoring how much processor power each program is consuming.

30 Days: Excellent for identifying long-term patterns and persistent problems.

7 Days: Good for finding repeating weekly trends.

24 Hours: Best for live troubleshooting or examining recent spikes.

Since your App Service Plan (P0V3, 2 instances) is reaching 100% CPU and 87% memory usage with each app consuming a maximum of 200MB and 40MB average, CPU-intensive activities or memory leaks might be occurring. Use Azure Monitor to contrast CPU and memory usage per application.

Check App Service Diagnostics for performance. Scale out through adding additional instances or reconfiguring high-use apps to their own plan in case bottlenecks occur.

https://learn.microsoft.com/en-us/azure/app-service/troubleshoot-diagnostic-logs
https://learn.microsoft.com/en-us/azure/app-service/web-sites-monitor
https://learn.microsoft.com/en-us/azure/app-service/monitor-app-service

If the answer is helpful, please don't forget to click Accept Answer and kindly upvote it and also let me know if you have any further Queries.
Bhargavi Naragani 1,690 Reputation points Microsoft External Staff

2025-03-07T23:11:57.3366667+00:00

Hi @ReCloudS,
Just checking in to see if above information was helpful. If you have any further updates on this issue, please feel free to post back.
ReCloudS 80 Reputation points

2025-03-09T06:16:50.36+00:00

@Bhargavi Naragani Sure am just check o what you suggested , will get back to you if I have any difficulties in understanding
ReCloudS 80 Reputation points

2025-03-10T04:28:51.9466667+00:00

@Bhargavi Naragani and also regarding the app service plan hosting multiple apps like 9-10 apps on a p2v3 plan with 3 instance count in that there were 5 web apps and 3 function apps when i check for sum of memory 400% and sometimes 1000% and cpu 50% to 300%(occasionally) in my app service plan for 30 days , here when I have to look in for the consumption in the individual app wise I don't see any problem meaning as the function apps were consuming pretty less compared to web apps so now can I reduce the instance count and highly consuming app to seperate app service plan where function apps consider them moving to the consumption will be cost optimised or making them stay in the same plan would make cost effective ? if all to look at the consumption of my existing function in cost wise can I see that ?
ReCloudS 80 Reputation points

2025-03-10T04:58:09.67+00:00

@Bhargavi Naragani , quick doubt on the consumption of the app service plan , as I have a plan of P2V3 with 3 instances which has 9 apps on it running which one web app consumes most of it and am considering it to move to the isolated app plan with required specs while I have 4 web apps and 4 function apps remaining so they were hardly consuming the cpu and memory as I made it in th earlier comment. Hence , do I need to make them to consumption plan for the function apps if yes what factors or metrics I need to consider for them to move to the conusmption from the premium app service plan whihc they are in now . Will they become cost optimised ? if I make them consumption and reduce the instance count or change the app service plan to a bit lower to P1V3 ?
Bhargavi Naragani 1,690 Reputation points Microsoft External Staff

2025-03-10T18:00:51.19+00:00
Hi @ReCloudS,
It definitely makes sense to relocate the web app that's consuming most of the CPU and memory to its own App Service Plan with the appropriate specs. This will ensure it won't affect the performance of your other apps and enable more efficient resource allocation.

Relocating the function apps to a Consumption Plan can certainly be cost-effective, but:

If your function apps run intermittently and don't have high execution times, the Consumption Plan is perfect since you only pay for real execution time.

The Consumption Plan dynamically allocates resources. If your function apps need steady or high performance, the Premium Plan (P1V3/P2V3) may still be better.

Consumption Plans scale automatically according to demand, which is perfect for variable workloads. If your function apps have stable usage, the existing plan might be more stable and predictable.

Functions in a Consumption Plan can suffer from cold starts, causing latency. If low-latency performance is a priority, you might prefer to remain with the Premium Plan.

Since the remaining 4 web apps and 4 function apps are consuming relatively little CPU and memory:

You could reduce the instance count from 3 to 2, especially if the summed usage remains below capacity.

Alternatively, downgrade to P1V3 if performance metrics over the last 30 days show that even with occasional spikes, you’re not consistently approaching resource limits.

https://azure.microsoft.com/en-us/pricing/details/app-service/linux/
https://learn.microsoft.com/en-us/azure/azure-functions/consumption-plan
https://learn.microsoft.com/en-us/azure/app-service/manage-scale-up
ReCloudS 80 Reputation points

2025-03-11T10:01:33.04+00:00

@Bhargavi Naragani , understood , only doubt left was as we have already deployed the function app on app service plan , I would like to know what are all the to consider for comparing to keep it on the app service plan or move to consumption plan what are all the metrics I need to consider in the Monitor>Metrics>Functionapp because I have considered function execution count only 7 avg/30 days and function execution units are 1.34M avg/month response time would be 543.71ms avg/30days are these enough to consider them moving to consumption inplace of having them in app service plan ? and may also expect cold start right ? or the response time would be sufficient
Bhargavi Naragani 1,690 Reputation points Microsoft External Staff

2025-03-11T11:00:24.8666667+00:00

@ReCloudS,

Based on what you've shared, here's how we analyze those numbers:

Function Execution Count: 7 avg/30 days — This is very low frequency, which makes the Consumption Plan perfect because you only get charged per execution.

Execution Units: 1.34M avg/month — This number represents total compute resources consumed. For the Consumption Plan, you get charged by execution time and memory, so this works perfectly with its pricing model.

Response Time: 543.71ms avg/30 days — This is acceptable but be aware that switching to the Consumption Plan adds the potential for cold starts, which can cause occasional latency spikes when the function isn’t warmed up. If your function apps need persistent low-latency response, this is something to consider.

Move to Consumption Plan if optimization of cost is the priority and occasional cold starts are acceptable. Remain on App Service Plan or switch to the Premium Plan if there is a requirement for consistent response time and performance without cold starts.

https://learn.microsoft.com/en-us/azure/azure-functions/functions-premium-plan?tabs=portal
https://learn.microsoft.com/en-us/azure/azure-functions/functions-best-practices?tabs=csharp

If the answer is helpful, please don't forget to click Accept Answer and kindly upvote it and also let me know if you have any further Queries.

Answer 2

You're on the right track in analyzing your Azure App Service Plan (ASP) metrics, and it does look like one of your web apps is consuming more memory, which is causing critical memory errors. You might want to consider the following:

Instance Utilization (~60-70%) – This suggests that, on average, each instance is operating at a moderate load.
Summed Utilization (~1000-1200%) – This sum is across all instances, meaning across the 3 instances, total usage is 10x-12x, confirming that resources are fully utilized.
Critical Memory Errors – Indicates that at least one app is hitting memory limits and potentially impacting the others.

1. Identify the app consuming the most memory
Use the App Service Metrics per app:

Go to Azure Monitor → Metrics → Select the Web App
Choose Memory Working Set
Set Aggregation Type: Avg and Max
Compare across all 6 apps to find the top consumer.

To mitigate:

Option 1: Isolate the high-consumption app

If a single web app is consuming excessive memory, consider moving it to a separate App Service Plan to prevent resource contention.
Deploy it on another P2V3 or higher-tier plan to distribute load effectively.

Option 2: Increase the instance count

Scaling out to more instances (e.g., from 3 to 4-5) will distribute load better, but it won't fix an app with a memory leak.
If the app is inefficient, scaling won't solve the core problem and may just spread the issue.

Option 3: Optimize the high-usage web app

Check Memory Profiler in Application Insights to diagnose high memory usage.
Optimize caching, GC settings, and reduce memory leaks in code.

To measure app-specific memory usage:

App Service Metrics → Memory Working Set (Per App)
Kudu (Advanced Tools) → Process Explorer
- Open Kudu (https://<yourapp>.scm.azurewebsites.net)
- Go to Process Explorer and check per-process memory consumption.

Effectively:

If one app is consuming excessive memory, move it to a separate App Service Plan
If all apps are moderately consuming memory, scale out by adding instances
If memory usage is abnormal, investigate memory leaks or optimize app performance

If the above response helps answer your question, remember to "Accept Answer" so that others in the community facing similar issues can easily find the solution. Your contribution is highly appreciated.

hth

Marcin

ReCloudS 80 Reputation points

2025-03-06T10:29:52.53+00:00

Thank you @Marcin Policht for the reply . I have followed the steps that you mentioned I have gone to the monitor and tried checking for the apps which are consuming more , I found out that one major app is consuming 9GB Max and average 3GB for 7 days and the other 5 apps were consuming 300MB Max and Avg lasting for 50MB here I need clarification because P2V3 has 16GB RAM with 3 Instance count which is 48GB in total , if it is consuming only max for 9GB and the other apps were hardly consuming 3-4 GB but for my understanding it shouldn't be a bottleneck right ? how come then it is having memory consumtion issue? and also the app service plan costs based on the usage right meaning if it costs 1000USD a month for 3 instance count it will be 3000 USD right ? and apps in and out data charges.

Share via

Multiple Web apps on a App service plan

1 additional answer

Your answer