The GDN Status Monitor Dashboard gives you visibility into how Smartling’s Global Delivery Network (GDN) proxy is performing for your localized websites. This guide explains the metrics shown on the Status Monitor pages, what the numbers mean, and how to interpret them for your specific setup.
Interpreting and evaluating metrics
There is no universal set of "good" numbers that applies to every account. What is considered healthy for a high-traffic enterprise site with dozens of domains may look very different from a smaller site serving only a few localized markets.
The health and speed of your source site are also critical factors and are not controlled by Smartling. Because the GDN is a proxy, it acts as an intermediary between your source site and the localized site. If your source site is slow, it will increase the total request time through the proxy.
Requests Last 24h
The total number of HTTP requests routed through the GDN proxy in the past 24 hours.
What it tells you: This is a baseline indicator of traffic volume. Large swings in either direction are worth investigating. A significant drop could indicate a configuration issue, a DNS problem, or reduced traffic. A large spike could indicate a traffic surge or a potential attack.
What is a good number: This metric is relative. A drop in requests is not inherently bad and may reflect normal variation, such as weekend or seasonal patterns. Establish your own baseline over time and investigate significant deviations rather than comparing against a universal benchmark.
AVG Total Request Time
The average time, in milliseconds, to fully process and respond to a request as measured by the GDN proxy. This includes both Proxy Overhead and Upstream Response.
What it tells you: This is the primary end-to-end latency metric for your localized pages. Rising average request times may indicate slower page loads for users.
What is a good number: This depends heavily on your origin server performance and audience geography. A well-optimized origin server may produce total request times under 100ms, while a slower or geographically distant origin server may see 300ms or more under normal conditions. Trends matter more than absolute values. If Proxy Overhead is stable but Upstream Response is increasing, the issue is likely on your origin server.
Proxy Overhead
The average time, in milliseconds, that Smartling's GDN proxy layer spends processing a request. This includes content retrieval, translation injection, and request routing.
What it tells you: This isolates latency introduced by the GDN itself and helps distinguish between proxy processing time and origin server performance.
What is a good number: Proxy Overhead is typically a small percentage of the total request time. For example, 19 ms out of an AVG Total Request Time of 245 ms is roughly 8 percent. A low and stable Proxy Overhead like this indicates efficient proxy processing. If it rises significantly and consistently, contact your Smartling Solution Architect or another Smartling representative.
Upstream Response
The time, in milliseconds, your origin server takes to respond to a GDN request. This measures how long your backend takes to deliver content before Smartling can process and serve it.
What it tells you: This reflects your own infrastructure's performance. If Upstream Response is high or increasing, the bottleneck is on your origin server, not within the GDN.
What is a good number: This depends entirely on your infrastructure. If total request time is high but Proxy Overhead is low, focus on optimizing your origin server rather than your GDN configuration.
HTTP Response Code Alerts (Last 24h)
This table breaks down requests from the past 24 hours by HTTP response code category. Each row shows a Total Count, a Percentage of overall traffic, and a Trend compared to the previous day.
The Trend indicator shows the percentage change compared to yesterday. A green indicator represents a favorable change and a red indicator represents an unfavorable change. Whether an increase or decrease is favorable depends on the response code category. For example, an increasing 2xx rate is positive, while an increasing 5xx rate is negative.
2xx Success
Requests that were successfully served. This is the most important row in the table.
What is a good number: For most sites, a 2xx rate in the high 90% range is healthy. A sudden or significant drop warrants investigation. A red trend indicator on this row means your success rate decreased and should be reviewed, especially if it is a large drop.
3xx Redirect
Requests that resulted in a redirect response. Some level of 3xx traffic is normal. Redirects are commonly used for HTTPS upgrades, canonical URL enforcement, and locale-based routing.
What is a good number: This depends on your site architecture. Stability is key, a sudden spike may indicate misconfigured redirect rules or a redirect loop.
4xx Client Error
Requests that resulted in a client-side error, most commonly 404 Not Found responses.
What is a good number: A small and stable percentage is normal. Crawlers and users occasionally request invalid URLs. A rising trend may indicate broken links, misconfigured GDN rules, or content that was removed without proper redirects in place.
5xx Server Error
Requests that resulted in a server-side error from your origin server.
What is a good number: A healthy 5xx rate should be as close to zero as possible. Even a small percentage means users are receiving error pages instead of content. Under normal conditions, most sites maintain a 5xx rate below 0.1 percent. A rising trend or sudden spike should be investigated promptly, as it usually indicates an issue with your origin server.
Reading the metrics together
The metrics on the Status Monitor pages are most useful when interpreted together rather than in isolation.
High AVG Total Request Time + Low Proxy Overhead + High Upstream Response: The GDN is performing normally and the latency is originating from your source site infrastructure.
Rising 4xx rate + Stable 2xx rate: You may have broken links or content that was removed without redirects. Review recent site changes.
Falling 2xx rate + Rising 5xx rate: Your origin server is returning errors. This may also appear as an Unhealthy status in the System Health section.
Large drop in Requests Last 24h + No change in error rates: Traffic may have dropped due to normal variation, a marketing change, or a DNS or routing issue upstream of the GDN.
When investigating issues, use the Trend column to identify what is changing and compare Proxy Overhead with Upstream Response to determine whether the issue originates within the GDN or in your own infrastructure.