Comment out timeout diagram while we figure out something that works …

…better with no animation.

Signed-off-by: Flynn <[email protected]>

linkerd.io/content/blog/2023/0912-linkerd-214.md

@@ -73,10 +73,26 @@ Additionally, this direct pod-to-pod communication reduces latency and also
 means that there are fewer components to fail in the network path. There's
 also no need for load balancers, which can directly cut costs.
 
-To use this new capability, you start with the `linkerd multicluster link`,
-just as you would have with earlier versions of Linkerd (and even though the
-gateways are no longer in play, you'll still see Link resources). However,
-when you label Services for export across clusters, you'll use
+To use this new capability, you start by installing the Linkerd multicluster
+extension with gateways disabled:
+
+```bash
+linkerd multicluster install --gateway=false | kubectl apply -f -
+```
+
+and then also add `--gateway=false` when linking clusters, e.g.:
+
+```bash
+linkerd --context=us-west multicluster link \
+        --cluster-name us-west \
+        --gateway=false \
+    | kubectl --context=us-east apply -f -
+```
+
+Note that even though the gateways are no longer in play, you'll still see the
+Link resources.
+
+Finally, when you label Services for export across clusters, you'll use
 
 ```yaml
 mirror.linkerd.io/exported: remote-discovery
@@ -157,10 +173,10 @@ The two types of timeouts have different purposes:
 - `timeouts.backendRequest` sets the timeout for just the communications from
   the proxies to the workload.
 
-{{< fig
+<!-- {{< fig
     alt="Linkerd 2.14 HTTPRoute Timeouts"
     title="Linkerd 2.14 HTTPRoute timeout semantics"
-    src="/uploads/2023/09/linkerd-2-14-httproute-timeouts.png" >}}
+    src="/uploads/2023/09/linkerd-2-14-httproute-timeouts.png" >}} -->
 
 Either timeout can be set alone, or both can be set together (in which case
 `timeouts.backendRequest` must be less than `timeouts.request`). To disable a