diff --git a/.github/workflows/ci-backend.yml b/.github/workflows/ci-backend.yml
index a1a38a751c6..f62f5879353 100644
--- a/.github/workflows/ci-backend.yml
+++ b/.github/workflows/ci-backend.yml
@@ -61,8 +61,10 @@ jobs:
                         # the dependencies more clear if we separated the backend/frontend
                         # code completely
                         # really we should ignore ee/frontend/** but dorny doesn't support that
+                        # - '!ee/frontend/**'
+                        # including the negated rule appears to work
+                        # but makes it always match because the checked file always isn't `ee/frontend/**` 🙈
                         - 'ee/**/*'
-                        - '!ee/frontend/**'
                         - 'posthog/**/*'
                         - 'bin/*.py'
                         - requirements.txt
diff --git a/frontend/src/lib/api.ts b/frontend/src/lib/api.ts
index 502946a9a07..854e5b657e4 100644
--- a/frontend/src/lib/api.ts
+++ b/frontend/src/lib/api.ts
@@ -1555,17 +1555,18 @@ const api = {
                 .withQueryString(toParams({ source: 'blob', blob_key: blobKey, version: '2' }))
                 .getResponse()
 
+            const contentBuffer = new Uint8Array(await response.arrayBuffer())
             try {
-                const textLines = await response.text()
+                const textDecoder = new TextDecoder()
+                const textLines = textDecoder.decode(contentBuffer)
 
                 if (textLines) {
                     return textLines.split('\n')
                 }
             } catch (e) {
-                // Must be gzipped
+                // we assume it is gzipped, swallow the error, and carry on below
             }
 
-            const contentBuffer = new Uint8Array(await response.arrayBuffer())
             return strFromU8(decompressSync(contentBuffer)).trim().split('\n')
         },
 
diff --git a/frontend/src/scenes/session-recordings/player/sessionRecordingDataLogic.test.ts b/frontend/src/scenes/session-recordings/player/sessionRecordingDataLogic.test.ts
index adf0e836d45..e75910bb916 100644
--- a/frontend/src/scenes/session-recordings/player/sessionRecordingDataLogic.test.ts
+++ b/frontend/src/scenes/session-recordings/player/sessionRecordingDataLogic.test.ts
@@ -127,6 +127,7 @@ describe('sessionRecordingDataLogic', () => {
                 })
             resumeKeaLoadersErrors()
         })
+
         it('fetch metadata success and snapshots error', async () => {
             silenceKeaLoadersErrors()
             // Unmount and remount the logic to trigger fetching the data again after the mock change
diff --git a/jest.setup.ts b/jest.setup.ts
index 847624fd022..cab389d3d46 100644
--- a/jest.setup.ts
+++ b/jest.setup.ts
@@ -1,6 +1,12 @@
 import 'whatwg-fetch'
 import 'jest-canvas-mock'
 
+import { TextEncoder, TextDecoder } from 'util'
+// Jest/JSDom don't know about TextEncoder but the browsers we support do
+// https://github.com/jsdom/jsdom/issues/2524
+global.TextDecoder = TextDecoder
+global.TextEncoder = TextEncoder
+
 window.scrollTo = jest.fn()
 window.matchMedia = jest.fn(
     () => ({ matches: false, addListener: jest.fn(), removeListener: jest.fn() } as MediaQueryList)