anomalyco/opencode
1
0
Fork 0
mirror of https://github.com/anomalyco/opencode.git synced 2026-07-21 17:56:33 -04:00
Code Issues 5.9k Packages Projects Releases 19 Wiki Activity

Big Pickle via OpenCode Zen is down this morning? [reporting similar to downdetector.com?] #6700

New Issue
Open
opened 2026-02-16 18:04:59 -05:00 by yindo · 1 comment
yindo commented 2026-02-16 18:04:59 -05:00
Owner
Copy Link
Copy Source

Originally created by @davidbernat on GitHub (Jan 18, 2026).

Originally assigned to: @fwang on GitHub.

Description

OpenCode team du excellence. The BigPickle model is clearly "down" in the traditional sense today. Latencies of 240+ seconds for basic tasks (i.e., take this console context and write to a file) and general failures spiked (i.e., </tool_call> to console log gibberish). The vision of OpenCode includes, one presumes, as we move towards models of the web that increase agent-interactions role as the former browsers of the Internet, provision and reporting of agent problems, similar to downdetector.com provides centralized crowdsourced reporting of website hosts (rip crowdsource x). This feature of the agentic AI would be immensely useful, and OpenCode has an opportunity to differentiate itself further as an open source model of agentic AI by establishing its principals.

Luckily, we have all the meta-reasoning required for such a tool directly in the workflows of OpenCode. (And, after last week finally cracking the Plugin model with Python yours truly can even contribute.) We know how long each request to the LLM takes for every user, and can easily apply very basic statistics to the anonymized data. That alone goes a long way, before considering pseudo-stats like whether users are sending duplicate prompts (an implicit failure) or console logs contain tool call crashes. Imagine how feature-ahead OpenCode investors could be if users like me could know (and confirm) ahead of time that Big Pickle (or any LLM) is shall we say taking the morning off?

You're missing the biggest value of all by simply collecting the metadata of these requests.
I left one major use case out of this discussion to see who bothers to respond. Can you spot it?
As for me, the V100 is not hooked up locally so I am cracking Genius Makers and reading instead.
Happy to lead this PR in Python.

Originally created by @davidbernat on GitHub (Jan 18, 2026). Originally assigned to: @fwang on GitHub. ### Description OpenCode team du excellence. The BigPickle model is clearly "down" in the traditional sense today. Latencies of 240+ seconds for basic tasks (i.e., take this console context and write to a file) and general failures spiked (i.e., </tool_call> to console log gibberish). The vision of OpenCode includes, one presumes, as we move towards models of the web that increase agent-interactions role as the former browsers of the Internet, provision and reporting of agent problems, similar to downdetector.com provides centralized crowdsourced reporting of website hosts (rip crowdsource x). This feature of the agentic AI would be immensely useful, and OpenCode has an opportunity to differentiate itself further as an open source model of agentic AI by establishing its principals. Luckily, we have all the meta-reasoning required for such a tool directly in the workflows of OpenCode. (And, after last week finally cracking the Plugin model with Python yours truly can even contribute.) We know how long each request to the LLM takes for every user, and can easily apply very basic statistics to the anonymized data. That alone goes a long way, before considering pseudo-stats like whether users are sending duplicate prompts (an implicit failure) or console logs contain tool call crashes. Imagine how feature-ahead OpenCode investors could be if users like me could know (and confirm) ahead of time that Big Pickle (or any LLM) is shall we say taking the morning off? You're missing the biggest value of all by simply collecting the metadata of these requests. I left one major use case out of this discussion to see who bothers to respond. Can you spot it? As for me, the V100 is not hooked up locally so I am cracking Genius Makers and reading instead. Happy to lead this PR in Python.
yindo added the bugzen labels 2026-02-16 18:04:59 -05:00
yindo commented 2026-02-16 18:05:00 -05:00
Author
Owner
Copy Link
Copy Source

@github-actions[bot] commented on GitHub (Jan 18, 2026):

This issue might be related to several existing issues discussing model performance monitoring and status reporting. Please check:

  • #9037: Add LLM API Call Duration Logging - foundational performance metrics collection
  • #6767: [Feature Request] Implement /stats command for visual usage insights - centralized stats monitoring similar to what you propose
  • #5555: [FEATURE]: Session-specific stats command and exit summary - performance statistics infrastructure
  • #9045: Codex model response latency increases significantly as conversation grows - actual model latency issues
  • #9172: [FEATURE]: Prompt + file edit logging for benchmarking - performance tracking and benchmarking
  • #6096: [FEATURE]: Adding Experimental Calculation and Display of Tokens per second - performance metrics display
  • #8193: experimental_telemetry metadata removed during LLM refactor - indicates telemetry infrastructure previously existed
  • #8322: feat: Show background tasks status in sidebar - status visibility for tasks
  • #7533: [FEATURE]: display loaded skills in status page and sidebar - status page functionality

These issues collectively address aspects of your feature request for centralized LLM performance reporting and monitoring. The infrastructure for collecting and displaying performance metadata has been requested across multiple issues.

Feel free to ignore if your specific case is different or if these don't fully address your needs.

@github-actions[bot] commented on GitHub (Jan 18, 2026): This issue might be related to several existing issues discussing model performance monitoring and status reporting. Please check: - #9037: Add LLM API Call Duration Logging - foundational performance metrics collection - #6767: [Feature Request] Implement /stats command for visual usage insights - centralized stats monitoring similar to what you propose - #5555: [FEATURE]: Session-specific stats command and exit summary - performance statistics infrastructure - #9045: Codex model response latency increases significantly as conversation grows - actual model latency issues - #9172: [FEATURE]: Prompt + file edit logging for benchmarking - performance tracking and benchmarking - #6096: [FEATURE]: Adding Experimental Calculation and Display of Tokens per second - performance metrics display - #8193: experimental_telemetry metadata removed during LLM refactor - indicates telemetry infrastructure previously existed - #8322: feat: Show background tasks status in sidebar - status visibility for tasks - #7533: [FEATURE]: display loaded skills in status page and sidebar - status page functionality These issues collectively address aspects of your feature request for centralized LLM performance reporting and monitoring. The infrastructure for collecting and displaying performance metadata has been requested across multiple issues. Feel free to ignore if your specific case is different or if these don't fully address your needs.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
acp
bad
bug
discussion
docs
documentation
enhancement
extension
good first issue
growth
help-wanted
mcp
model-problem
needs:compliance
nix
opentui
perf
pull-request
Mirrored from GitHub Pull Request
 tui
web
windows
zen
No Label bug zen
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: anomalyco/opencode#6700
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?