self-hosted icon indicating copy to clipboard operation
self-hosted copied to clipboard

unable see the Trace details

Open Krishpluto opened this issue 5 months ago • 19 comments

Self-Hosted Version

25.5.1

CPU Architecture

x86_64

Docker Version

25.0.3

Docker Compose Version

2.32.2

Machine Specification

  • [x] My system meets the minimum system requirements of Sentry

Steps to Reproduce

  1. Navigate to the sentry project dashboard
  2. Go to any specific issue
  3. If I am in the issue view in sentry and click on “View Full Trace" link
  4. observe the error message displayed instead of the trace details.

We have verified this behavior across multiple projects, including a dedicated test project, and have found that even admin users are unable to view it. We have confirmed that the feature is enabled on our end.

Expected Result

We expected the full trace view to open and display the transaction trace timeline and related performance data.

Actual Result

Instead of showing the trace, the page returns the error message:

"You don't have access to this feature"

Image

Event ID

No response

Krishpluto avatar Jun 03 '25 11:06 Krishpluto

  1. Can you show me what's your feature flags on your sentry.conf.py?
  2. Have you made sure that the value of version tag on your .env and/or .env.custom is 25.5.1?

aldy505 avatar Jun 05 '25 02:06 aldy505

Hi @aldy505 ,

Thanks for the reply. Please see the below list of features that we have enabled from our end.

Image

Regarding the .env and/or .env.custom can you tell me where to check this variable in the code file? But upon checking from our end, we are in the current version in the Sentry

Krishpluto avatar Jun 05 '25 08:06 Krishpluto

I have the exact same issue on 25.4.0.

ENVs are configured correctly. Do I need to change anything to enable this feature after the update? Currently my account is in the admin role and I should be able to see everything(and previously it worked fine), but traces are not available now:

You don't have access to this feature

Romanavr avatar Jun 09 '25 12:06 Romanavr

Hello @Krishpluto, I see you are missing a lot of feature flags. Please refer to this file here: https://github.com/getsentry/self-hosted/blob/1217f469ec4a7d1f8329494f7b61927f5156c28a/sentry/sentry.conf.example.py#L262-L313

The .env file can be checked here: https://github.com/getsentry/self-hosted/blob/25.5.1/.env you should have it pointed to 25.5.1 at the moment, as it is a hard stop.

Also just to make sure, try to see the differences between your sentry.conf.py files by doing: diff sentry/sentry.conf.py sentry/sentry.conf.example.py. You'll need to resolve that and make sure you have COMPOSE_PROFILES environment variable set to feature-complete and make sure this line exists on your sentry.conf.py: https://github.com/getsentry/self-hosted/blob/cbebc4f3f43a6ef5834d059bb144e67b7515d752/sentry/sentry.conf.example.py#L71-L81

aldy505 avatar Jun 09 '25 12:06 aldy505

I have updated the config to have all the features from the example config, but I still get the same “You do not have access to this feature” Any ideas @aldy505 ?

Romanavr avatar Jun 09 '25 18:06 Romanavr

I have the same issue on 25.5.1. .env and sentry.conf.py files are up to date with examples from github.

pgodzwa avatar Jun 10 '25 07:06 pgodzwa

I have updated the config to have all the features from the example config, but I still get the same “You do not have access to this feature” Any ideas @aldy505 ?

@Romanavr have you re-run ./install.sh yet? What did you do after saving the configuration file with the new data?

aldy505 avatar Jun 10 '25 07:06 aldy505

Hi @aldy505 ,

We are setting up a self-hosted Sentry instance for a client who has specifically requested support for trace-related features only—not the full feature set.

I’ve reviewed the sentry.conf.example.py and .env configurations, and I understand that enabling COMPOSE_PROFILES=feature-complete activates all features. However, to minimize resource usage, we’d like to enable only the essential features required for distributed tracing (e.g., trace view, span ingestion, dynamic sampling, etc.).

Could you please confirm which specific feature flags from SENTRY_FEATURES.update() are strictly required to support tracing functionality?

Krishpluto avatar Jun 10 '25 08:06 Krishpluto

Hi @aldy505 ,

We are setting up a self-hosted Sentry instance for a client who has specifically requested support for trace-related features only—not the full feature set.

I’ve reviewed the sentry.conf.example.py and .env configurations, and I understand that enabling COMPOSE_PROFILES=feature-complete activates all features. However, to minimize resource usage, we’d like to enable only the essential features required for distributed tracing (e.g., trace view, span ingestion, dynamic sampling, etc.).

Could you please confirm which specific feature flags from SENTRY_FEATURES.update() are strictly required to support tracing functionality?

@Krishpluto Try this, set your COMPOSE_PROFILES to feature-complete. Then on your sentry.conf.py:

SENTRY_FEATURES.update(
    {
        feature: True
        for feature in (
            "organizations:discover",
            "organizations:global-views",
            "organizations:incidents",
            "organizations:integrations-issue-basic",
            "organizations:integrations-issue-sync",
            "organizations:invite-members",
            "organizations:sso-basic",
            "organizations:sso-rippling",
            "organizations:sso-saml2",
            "organizations:performance-view",
            "organizations:advanced-search",
            "organizations:session-replay",
            "organizations:issue-platform",
            "organizations:dashboards-mep",
            "organizations:mep-rollout-flag",
            "organizations:dashboards-rh-widget",
            "organizations:metrics-extraction",
            "organizations:transaction-metrics-extraction",
            "organizations:trace-view-v1",
            "organizations:dynamic-sampling",
            "projects:custom-inbound-filters",
            "projects:data-forwarding",
            "projects:discard-groups",
            "projects:plugins",
            "projects:rate-limits",
            "projects:servicehooks",
        )
        # Starfish related flags
        + (
            "organizations:indexed-spans-extraction",
            "organizations:insights-entry-points",
            "organizations:insights-initial-modules",
            "organizations:insights-addon-modules",
            "organizations:standalone-span-ingestion",
            "organizations:starfish-mobile-appstart",
            "projects:span-metrics-extraction",
            "projects:span-metrics-extraction-addons",
        )
    }
)

On your docker-compose.yml, you can remove a few containers (these are the containers that you can remove):

  • https://github.com/getsentry/self-hosted/blob/1217f469ec4a7d1f8329494f7b61927f5156c28a/docker-compose.yml#L247-L251
  • https://github.com/getsentry/self-hosted/blob/1217f469ec4a7d1f8329494f7b61927f5156c28a/docker-compose.yml#L292-L306
  • https://github.com/getsentry/self-hosted/blob/1217f469ec4a7d1f8329494f7b61927f5156c28a/docker-compose.yml#L397-L421
  • https://github.com/getsentry/self-hosted/blob/1217f469ec4a7d1f8329494f7b61927f5156c28a/docker-compose.yml#L490-L521

After that, you can re-run ./install.sh script. The only feature that you see now is errors & tracing. Everything else (crons, session replay, and profiling) are removed.

aldy505 avatar Jun 10 '25 08:06 aldy505

I have updated the config to have all the features from the example config, but I still get the same “You do not have access to this feature” Any ideas @aldy505 ?

@Romanavr have you re-run ./install.sh yet? What did you do after saving the configuration file with the new data?

Yes, I ran install.sh and docker compose with my custom .ENV file after changing the sentry.config.py file. Is there anything else I need to do?

Romanavr avatar Jun 10 '25 09:06 Romanavr

I have updated the config to have all the features from the example config, but I still get the same “You do not have access to this feature” Any ideas @aldy505 ?

@Romanavr have you re-run ./install.sh yet? What did you do after saving the configuration file with the new data?

Yes, I ran install.sh and docker compose with my custom .ENV file after changing the sentry.config.py file. Is there anything else I need to do?

@Romanavr Can you show me your .env and sentry.conf.py with the secrets redacted?

aldy505 avatar Jun 11 '25 01:06 aldy505

After that, you can re-run ./install.sh script. The only feature that you see now is errors & tracing. Everything else (crons, session replay, and profiling) are removed.

Hi @aldy505

Apologies for the earlier lack of clarity in my communication.

We currently have several features enabled—such as crons and session replay—which we would like to retain as-is. Our intention is solely to incorporate the trace functionality to address the current issue, without impacting the existing setup.

Could you please advise on which containers need to be added or modified to enable tracing, while ensuring that the existing features remain unaffected?

Krishpluto avatar Jun 11 '25 14:06 Krishpluto

After that, you can re-run ./install.sh script. The only feature that you see now is errors & tracing. Everything else (crons, session replay, and profiling) are removed.

Hi @aldy505

Apologies for the earlier lack of clarity in my communication.

We currently have several features enabled—such as crons and session replay—which we would like to retain as-is. Our intention is solely to incorporate the trace functionality to address the current issue, without impacting the existing setup.

Could you please advise on which containers need to be added or modified to enable tracing, while ensuring that the existing features remain unaffected?

@Krishpluto you would only need to enable the "trace-view-v1" feature flag from the feature flag list. Again, see my comment here: https://github.com/getsentry/self-hosted/issues/3729#issuecomment-2955697601

The excess features would be profiling and user feedback, which doesn't really consume that much resources.

aldy505 avatar Jun 12 '25 02:06 aldy505

I am having the same issue with traces, "You don't have access to this feature" ...

using .env included with project & sentry.conf.py is generated by install.sh ... only customization to sentry.conf.py is the addition of CSRF_TRUSTED_ORIGINS

current sentry version, Sentry 25.5.1

alloylab avatar Jun 12 '25 12:06 alloylab

I am having the same issue with traces, "You don't have access to this feature" ...

using .env included with project & sentry.conf.py is generated by install.sh ... only customization to sentry.conf.py is the addition of CSRF_TRUSTED_ORIGINS

current sentry version, Sentry 25.5.1

@alloylab The ./install.sh script won't modify any configuration files (that includes sentry.conf.py and .env). Please refer to my comment here: https://github.com/getsentry/self-hosted/issues/3729#issuecomment-2955697601

aldy505 avatar Jun 12 '25 12:06 aldy505

@aldy505 senty.conf.py is generated by install.sh ... see this install file in the project, install/ensure-files-from-examples.sh ; its only generated if the file doesn't exist

alloylab avatar Jun 12 '25 12:06 alloylab

@aldy505 senty.conf.py is generated by install.sh ... see this install file in the project, install/ensure-files-from-examples.sh ; its only generated if the file doesn't exist

@alloylab Yes, what I'm referring to is that: if there are any updates to the sentry.conf.example.py, your non-example Django config file (sentry.conf.py) won't get modified. You will need to update it yourself. Now, can you compare your sentry.conf.py to sentry.conf.example.py to see if there are any missing things?

aldy505 avatar Jun 12 '25 22:06 aldy505

@aldy505 I let the install.sh generate a fresh sentry.conf.py and I am still getting the error, "You don't have access to this feature" here is sentry.conf.py that install.sh generated... with the exception of line 404 that I add once the file is generated

sentry.conf.py

the only features missing are the ones you adde last week in https://github.com/getsentry/self-hosted/pull/3742/commits/f46ab3be2ff5829c7e7a653048f0accb5c8638a0

I will try updating again with the latest updates from last week and report back

alloylab avatar Jun 17 '25 16:06 alloylab

issue exists in master branch as well see attached sentry.conf.py

sentry.conf.py.txt

alloylab avatar Jun 17 '25 17:06 alloylab

Same for me, fresh install today but trace is not available in the menu and when i manually go to the page i get the message You don't have access to this feature aswell.

joey01245 avatar Jun 19 '25 21:06 joey01245

Okay people, let's debug this.

When you see "You don't have access to this feature", can you paste in the URL you see on the browser URL here? Let's see what feature flags are guarding the page.

aldy505 avatar Jun 20 '25 01:06 aldy505

Both of these urls are not working: https://sentry.xxx.nl/organizations/xxx/explore/traces https://sentry.xxx.nl/organizations/xxx/traces

joey01245 avatar Jun 20 '25 07:06 joey01245

Both of these urls are not working: https://sentry.xxx.nl/organizations/xxx/explore/traces https://sentry.xxx.nl/organizations/xxx/traces

@joey01245 Those paths requires the "organizations:trace-view-v1" feature flag being set on your sentry.conf.py. Have you set it?

aldy505 avatar Jun 20 '25 07:06 aldy505

Yes, i installed it fresh yesterday and it is in by default.

SENTRY_FEATURES["projects:sample-events"] = False
SENTRY_FEATURES.update(
    {
        feature: True
        for feature in (
            "organizations:discover",
            "organizations:global-views",
            "organizations:incidents",
            "organizations:integrations-issue-basic",
            "organizations:integrations-issue-sync",
            "organizations:invite-members",
            "organizations:sso-basic",
            "organizations:sso-rippling",
            "organizations:sso-saml2",
            "organizations:performance-view",
            "organizations:advanced-search",
            "organizations:session-replay",
            "organizations:issue-platform",
            "organizations:profiling",
            "organizations:monitors",
            "organizations:dashboards-mep",
            "organizations:mep-rollout-flag",
            "organizations:dashboards-rh-widget",
            "organizations:metrics-extraction",
            "organizations:transaction-metrics-extraction",
            "organizations:trace-view-v1",
            "organizations:dynamic-sampling",
            "projects:custom-inbound-filters",
            "projects:data-forwarding",
            "projects:discard-groups",
            "projects:plugins",
            "projects:rate-limits",
            "projects:servicehooks",
        )
        # Starfish related flags
        + (
            "organizations:indexed-spans-extraction",
            "organizations:insights-entry-points",
            "organizations:insights-initial-modules",
            "organizations:insights-addon-modules",
            "organizations:standalone-span-ingestion",
            "organizations:starfish-mobile-appstart",
            "projects:span-metrics-extraction",
            "projects:span-metrics-extraction-addons",
        )
        # User Feedback related flags
        + (
            "organizations:user-feedback-ingest",
            "organizations:user-feedback-replay-clip",
            "organizations:user-feedback-ui",
        )
        # Continuous Profiling related flags
        + (
            "organizations:continuous-profiling",
            "organizations:continuous-profiling-stats",
        )
    }
)

joey01245 avatar Jun 20 '25 07:06 joey01245

Image

Hi @aldy505 ,

"organizations:events organizations:metric-alert-builder-aggregate"

We have observed that the above mentioned feature flags have been officially deprecated in version 25.5.1. However, these flags are still present in our codebase. Could you please confirm if removing these flags will impact our current functionality?

Krishpluto avatar Jun 24 '25 06:06 Krishpluto

Hi @aldy505 ,

"organizations:events organizations:metric-alert-builder-aggregate"

We have observed that the above mentioned feature flags have been officially deprecated in version 25.5.1. However, these flags are still present in our codebase. Could you please confirm if removing these flags will impact our current functionality?

@Krishpluto it does not matter if a feature flag is present on the list, yet it's removed.

aldy505 avatar Jun 26 '25 04:06 aldy505

Yes, i installed it fresh yesterday and it is in by default.

SENTRY_FEATURES["projects:sample-events"] = False
SENTRY_FEATURES.update(
    {
        feature: True
        for feature in (
            "organizations:discover",
            "organizations:global-views",
            "organizations:incidents",
            "organizations:integrations-issue-basic",
            "organizations:integrations-issue-sync",
            "organizations:invite-members",
            "organizations:sso-basic",
            "organizations:sso-rippling",
            "organizations:sso-saml2",
            "organizations:performance-view",
            "organizations:advanced-search",
            "organizations:session-replay",
            "organizations:issue-platform",
            "organizations:profiling",
            "organizations:monitors",
            "organizations:dashboards-mep",
            "organizations:mep-rollout-flag",
            "organizations:dashboards-rh-widget",
            "organizations:metrics-extraction",
            "organizations:transaction-metrics-extraction",
            "organizations:trace-view-v1",
            "organizations:dynamic-sampling",
            "projects:custom-inbound-filters",
            "projects:data-forwarding",
            "projects:discard-groups",
            "projects:plugins",
            "projects:rate-limits",
            "projects:servicehooks",
        )
        # Starfish related flags
        + (
            "organizations:indexed-spans-extraction",
            "organizations:insights-entry-points",
            "organizations:insights-initial-modules",
            "organizations:insights-addon-modules",
            "organizations:standalone-span-ingestion",
            "organizations:starfish-mobile-appstart",
            "projects:span-metrics-extraction",
            "projects:span-metrics-extraction-addons",
        )
        # User Feedback related flags
        + (
            "organizations:user-feedback-ingest",
            "organizations:user-feedback-replay-clip",
            "organizations:user-feedback-ui",
        )
        # Continuous Profiling related flags
        + (
            "organizations:continuous-profiling",
            "organizations:continuous-profiling-stats",
        )
    }
)

@joey01245 Try to add organizations:performance-trace-details feature flag if you're on 25.6.x release, and see if that fixes it.

aldy505 avatar Jun 26 '25 04:06 aldy505

Thanks @aldy505 this is working! The only thing is that there is no sorting for the timestamps, in another issue i found organizations:performance-trace-explorer-sorting but this seems to be removed or not working, is there a solution for this?

joey01245 avatar Jun 26 '25 06:06 joey01245

Thanks @aldy505 this is working! The only thing is that there is no sorting for the timestamps, in another issue i found organizations:performance-trace-explorer-sorting but this seems to be removed or not working, is there a solution for this?

@joey01245 I believe that requires organizations:performance-trace-explorer to be enabled, but in order to enable that, you'll need to enable all EAP (event analytics platform) containers, which accumulates to around 6 new Docker containers.

My suggestion is to let it be, and wait for 25.7.0 release where we make another breaking change.

aldy505 avatar Jun 26 '25 06:06 aldy505

Alright, no problem, we will wait, thanks for the help!

joey01245 avatar Jun 26 '25 09:06 joey01245