I am currently experiencing an intermittent connectivity problem with my website, where the site loads normally for some visitors but becomes completely unreachable for others, depending on the network they are using. The website itself appears to be functioning correctly most of the time, and I can access it without any difficulty from my usual internet connection. However, I have received reports from visitors who cannot open the website at certain times, even though other websites remain accessible through their connections. In some cases, the page eventually loads after several attempts, while in other cases the browser displays a connection timeout or indicates that the server could not be reached. What makes this issue confusing is that the problem does not appear to affect everyone simultaneously, and I have not been able to identify a consistent pattern based on the time of day or the visitor’s browser. I want to understand whether this behavior could be related to internet routing or network reachability rather than a problem with the website’s application itself.
To investigate the issue, I have checked the website from multiple devices and internet connections, including a separate mobile connection, and the results have not always been consistent. There are periods when the website opens immediately from every connection I test, followed by periods when a particular network cannot reach it while another connection continues to work normally. I have also checked whether the website’s server is running during these incidents, and the available server information does not indicate a complete outage. The web server continues to respond to requests from at least some locations, which makes me wonder whether traffic from certain networks is taking an unsuccessful route or encountering a connectivity problem somewhere between the visitor and the hosting network. I understand that different networks can use different paths to reach the same destination, but I am not sure how to determine whether the observed behavior is actually caused by routing differences. I would appreciate guidance on how to distinguish a network-path problem from an ordinary hosting or connectivity issue without drawing conclusions from a few isolated tests.
I have reviewed the basic configuration on my side to make sure that the problem is not caused by an obvious mistake in the website’s setup. The domain resolves to the expected server address when I check it from my own connection, and I have not intentionally changed the hosting provider, server address, or website configuration around the time I first noticed the issue. I have also checked the web server logs, but they do not provide a clear explanation for requests that never appear to reach the server. This is particularly difficult to investigate because a failed connection does not necessarily produce an application-level error that I can inspect, especially when the request appears to stop before the website begins processing it. I do not want to assume that the issue is related to BGP or a particular upstream provider without sufficient evidence, but the fact that the website remains accessible from some networks while failing from others makes network-level diagnostics seem worth exploring. I am interested in understanding which measurements would provide reliable evidence about where the connection is failing.
Since RIPE provides tools and resources for examining internet connectivity and routing, I would like to know whether RIPEstat or RIPE Atlas could help me investigate this specific situation. For example, would it be useful to run measurements from multiple RIPE Atlas probes toward the website’s IP address while the problem is occurring, and compare the results with measurements taken when the site is accessible? I am particularly interested in identifying whether probes in certain networks consistently experience timeouts, unexpected routing changes, or a different network path from probes that can reach the destination successfully. I am also unsure how to interpret traceroute results when some intermediate routers do not respond, because I understand that routers may filter or deprioritize diagnostic packets even when ordinary traffic can pass through them. If anyone has experience using RIPE Atlas for this kind of intermittent reachability investigation, I would appreciate suggestions about suitable measurement types, probe selection, and the limitations I should consider before interpreting the results.
Another detail I am trying to understand is whether the origin network’s routing announcements or upstream connectivity could contribute to a problem that appears only from certain networks. I do not currently have enough evidence to say that a route is being withdrawn, that a prefix is being filtered, or that a routing policy is responsible, and I would prefer to verify these possibilities before contacting my hosting provider with an unsupported explanation. Would comparing the relevant prefix and origin information in RIPEstat, reviewing available BGP visibility, and collecting timestamps of successful and failed connection attempts be a reasonable starting point? If the destination prefix appears to be announced normally throughout the incident, would that rule out a routing-related problem, or could the issue still involve a more specific path or forwarding failure? I would also like to know which information would be most useful to collect before escalating the matter to the hosting provider or network operator, such as the affected source networks, destination IP address, timestamps, traceroute outputs, and whether TCP connections to the web server’s port succeed. My main goal is to gather reproducible evidence instead of relying only on reports that the website sometimes does not load.
Has anyone in the RIPE community encountered a similar situation in which a website remained available from some networks but became intermittently unreachable from others, without an obvious server outage? I would be grateful for advice on a practical troubleshooting sequence that would help narrow down the failure point and establish whether the problem lies within the hosting network, an upstream provider, or a remote network’s path toward the destination. In particular, I would like to know how to correlate RIPE Atlas measurements with BGP routing information and ordinary connection tests, and how much evidence is generally needed before a routing issue can be reported meaningfully to a network operator. If there are specific RIPE tools, measurement configurations, or diagnostic practices that are especially useful for this type of investigation, please share your recommendations. I am trying to focus on this single reachability issue and determine why access to the same website differs between networks, rather than making unnecessary changes to the server while the actual cause remains unclear. Thanks!