Revert "report macmon failures more aggressively (#1618)" (#1625)

ba611f9cd0e21d3e63e2327b18fbc888fd085269 · 2026-02-25 19:15:55 +0000 · Evan Quiney

this pr broke macmon - revert it

Files touched

M src/exo/utils/info_gatherer/info_gatherer.py

Diff

commit ba611f9cd0e21d3e63e2327b18fbc888fd085269
Author: Evan Quiney <evanev7@gmail.com>
Date:   Wed Feb 25 19:15:55 2026 +0000

    Revert "report macmon failures more aggressively (#1618)" (#1625)
    
    this pr broke macmon - revert it
---
 src/exo/utils/info_gatherer/info_gatherer.py | 19 ++++++-------------
 1 file changed, 6 insertions(+), 13 deletions(-)

diff --git a/src/exo/utils/info_gatherer/info_gatherer.py b/src/exo/utils/info_gatherer/info_gatherer.py
index 1efb1a90..9cb04c06 100644
--- a/src/exo/utils/info_gatherer/info_gatherer.py
+++ b/src/exo/utils/info_gatherer/info_gatherer.py
@@ -542,13 +542,10 @@ class InfoGatherer:
                     if not p.stdout:
                         logger.critical("MacMon closed stdout")
                         return
-                    t = TextReceiveStream(BufferedByteReceiveStream(p.stdout))
-                    while True:
-                        with anyio.fail_after(self.macmon_interval * 3):
-                            macmon_output = await t.receive()
-                            await self.info_sender.send(
-                                MacmonMetrics.from_raw_json(macmon_output)
-                            )
+                    async for text in TextReceiveStream(
+                        BufferedByteReceiveStream(p.stdout)
+                    ):
+                        await self.info_sender.send(MacmonMetrics.from_raw_json(text))
             except CalledProcessError as e:
                 stderr_msg = "no stderr"
                 stderr_output = cast(bytes | str | None, e.stderr)
@@ -559,12 +556,8 @@ class InfoGatherer:
                         else str(stderr_output)
                     )
                 logger.warning(
-                    f"memory monitor failed with return code {e.returncode}: {stderr_msg}"
-                )
-            except TimeoutError:
-                logger.warning(
-                    f"memory monitor silent for {self.macmon_interval * 3}s - reloading"
+                    f"MacMon failed with return code {e.returncode}: {stderr_msg}"
                 )
             except Exception as e:
-                logger.opt(exception=e).warning("Error in memory monitor")
+                logger.warning(f"Error in macmon monitor: {e}")
             await anyio.sleep(self.macmon_interval)

← eab3e0b4 report macmon failures more aggressively (#1618) · back to Exo · Use tmpdir for coordination file (#1624) eaed9295 →